1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
3 @c This is part of the GCC manual.
4 @c For copying conditions, see the file gcc.texi.
8 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
9 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
11 Permission is granted to copy, distribute and/or modify this document
12 under the terms of the GNU Free Documentation License, Version 1.2 or
13 any later version published by the Free Software Foundation; with the
14 Invariant Sections being ``GNU General Public License'' and ``Funding
15 Free Software'', the Front-Cover texts being (a) (see below), and with
16 the Back-Cover Texts being (b) (see below). A copy of the license is
17 included in the gfdl(7) man page.
19 (a) The FSF's Front-Cover Text is:
23 (b) The FSF's Back-Cover Text is:
25 You have freedom to copy and modify this GNU Manual, like GNU
26 software. Copies published by the Free Software Foundation raise
27 funds for GNU development.
29 @c Set file name and title for the man page.
31 @settitle GNU project C and C++ compiler
33 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
34 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
35 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
36 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
37 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
38 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
39 [@option{-o} @var{outfile}] @var{infile}@dots{}
41 Only the most useful options are listed here; see below for the
42 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
45 gpl(7), gfdl(7), fsf-funding(7),
46 cpp(1), gcov(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
47 and the Info entries for @file{gcc}, @file{cpp}, @file{as},
48 @file{ld}, @file{binutils} and @file{gdb}.
51 For instructions on reporting bugs, see
52 @w{@uref{http://gcc.gnu.org/bugs.html}}. Use of the @command{gccbug}
53 script to report bugs is recommended.
56 See the Info entry for @command{gcc}, or
57 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
58 for contributors to GCC@.
63 @chapter GCC Command Options
64 @cindex GCC command options
65 @cindex command options
66 @cindex options, GCC command
68 @c man begin DESCRIPTION
69 When you invoke GCC, it normally does preprocessing, compilation,
70 assembly and linking. The ``overall options'' allow you to stop this
71 process at an intermediate stage. For example, the @option{-c} option
72 says not to run the linker. Then the output consists of object files
73 output by the assembler.
75 Other options are passed on to one stage of processing. Some options
76 control the preprocessor and others the compiler itself. Yet other
77 options control the assembler and linker; most of these are not
78 documented here, since you rarely need to use any of them.
80 @cindex C compilation options
81 Most of the command line options that you can use with GCC are useful
82 for C programs; when an option is only useful with another language
83 (usually C++), the explanation says so explicitly. If the description
84 for a particular option does not mention a source language, you can use
85 that option with all supported languages.
87 @cindex C++ compilation options
88 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
89 options for compiling C++ programs.
91 @cindex grouping options
92 @cindex options, grouping
93 The @command{gcc} program accepts options and file names as operands. Many
94 options have multi-letter names; therefore multiple single-letter options
95 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
98 @cindex order of options
99 @cindex options, order
100 You can mix options and other arguments. For the most part, the order
101 you use doesn't matter. Order does matter when you use several options
102 of the same kind; for example, if you specify @option{-L} more than once,
103 the directories are searched in the order specified.
105 Many options have long names starting with @samp{-f} or with
106 @samp{-W}---for example, @option{-fforce-mem},
107 @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of
108 these have both positive and negative forms; the negative form of
109 @option{-ffoo} would be @option{-fno-foo}. This manual documents
110 only one of these two forms, whichever one is not the default.
114 @xref{Option Index}, for an index to GCC's options.
117 * Option Summary:: Brief list of all options, without explanations.
118 * Overall Options:: Controlling the kind of output:
119 an executable, object files, assembler files,
120 or preprocessed source.
121 * Invoking G++:: Compiling C++ programs.
122 * C Dialect Options:: Controlling the variant of C language compiled.
123 * C++ Dialect Options:: Variations on C++.
124 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
126 * Language Independent Options:: Controlling how diagnostics should be
128 * Warning Options:: How picky should the compiler be?
129 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
130 * Optimize Options:: How much optimization?
131 * Preprocessor Options:: Controlling header files and macro definitions.
132 Also, getting dependency information for Make.
133 * Assembler Options:: Passing options to the assembler.
134 * Link Options:: Specifying libraries and so on.
135 * Directory Options:: Where to find header files and libraries.
136 Where to find the compiler executable files.
137 * Spec Files:: How to pass switches to sub-processes.
138 * Target Options:: Running a cross-compiler, or an old version of GCC.
139 * Submodel Options:: Specifying minor hardware or convention variations,
140 such as 68010 vs 68020.
141 * Code Gen Options:: Specifying conventions for function calls, data layout
143 * Environment Variables:: Env vars that affect GCC.
144 * Precompiled Headers:: Compiling a header once, and using it many times.
145 * Running Protoize:: Automatically adding or removing function prototypes.
151 @section Option Summary
153 Here is a summary of all the options, grouped by type. Explanations are
154 in the following sections.
157 @item Overall Options
158 @xref{Overall Options,,Options Controlling the Kind of Output}.
159 @gccoptlist{-c -S -E -o @var{file} -combine -pipe -pass-exit-codes @gol
160 -x @var{language} -v -### --help --target-help --version}
162 @item C Language Options
163 @xref{C Dialect Options,,Options Controlling C Dialect}.
164 @gccoptlist{-ansi -std=@var{standard} -aux-info @var{filename} @gol
165 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
166 -fhosted -ffreestanding -fms-extensions @gol
167 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
168 -fallow-single-precision -fcond-mismatch @gol
169 -fsigned-bitfields -fsigned-char @gol
170 -funsigned-bitfields -funsigned-char}
172 @item C++ Language Options
173 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
174 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
175 -fconserve-space -fno-const-strings @gol
176 -fno-elide-constructors @gol
177 -fno-enforce-eh-specs @gol
178 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
179 -fno-implicit-templates @gol
180 -fno-implicit-inline-templates @gol
181 -fno-implement-inlines -fms-extensions @gol
182 -fno-nonansi-builtins -fno-operator-names @gol
183 -fno-optional-diags -fpermissive @gol
184 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
185 -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
186 -fno-default-inline -fvisibility-inlines-hidden @gol
187 -Wabi -Wctor-dtor-privacy @gol
188 -Wnon-virtual-dtor -Wreorder @gol
189 -Weffc++ -Wno-deprecated @gol
190 -Wno-non-template-friend -Wold-style-cast @gol
191 -Woverloaded-virtual -Wno-pmf-conversions @gol
192 -Wsign-promo -Wsynth}
194 @item Objective-C and Objective-C++ Language Options
195 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
196 Objective-C and Objective-C++ Dialects}.
198 -fconstant-string-class=@var{class-name} @gol
199 -fgnu-runtime -fnext-runtime @gol
200 -fno-nil-receivers @gol
201 -fobjc-exceptions @gol
202 -freplace-objc-classes @gol
205 -Wno-protocol -Wselector -Wundeclared-selector}
207 @item Language Independent Options
208 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
209 @gccoptlist{-fmessage-length=@var{n} @gol
210 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}}
212 @item Warning Options
213 @xref{Warning Options,,Options to Request or Suppress Warnings}.
214 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
215 -w -Wextra -Wall -Waggregate-return @gol
216 -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol
217 -Wconversion -Wno-deprecated-declarations @gol
218 -Wdisabled-optimization -Wno-div-by-zero -Wendif-labels @gol
219 -Werror -Werror-implicit-function-declaration @gol
220 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
221 -Wno-format-extra-args -Wformat-nonliteral @gol
222 -Wformat-security -Wformat-y2k @gol
223 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
224 -Wimport -Wno-import -Winit-self -Winline @gol
225 -Wno-invalid-offsetof -Winvalid-pch @gol
226 -Wlarger-than-@var{len} -Wlong-long @gol
227 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
228 -Wmissing-format-attribute -Wmissing-include-dirs @gol
229 -Wmissing-noreturn @gol
230 -Wno-multichar -Wnonnull -Wpacked -Wpadded @gol
231 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
232 -Wreturn-type -Wsequence-point -Wshadow @gol
233 -Wsign-compare -Wstrict-aliasing -Wstrict-aliasing=2 @gol
234 -Wswitch -Wswitch-default -Wswitch-enum @gol
235 -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol
236 -Wunknown-pragmas -Wunreachable-code @gol
237 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
238 -Wunused-value -Wunused-variable -Wwrite-strings @gol
241 @item C-only Warning Options
242 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
243 -Wmissing-prototypes -Wnested-externs -Wold-style-definition @gol
244 -Wstrict-prototypes -Wtraditional @gol
245 -Wdeclaration-after-statement -Wno-pointer-sign}
247 @item Debugging Options
248 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
249 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
250 -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
251 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
252 -fdump-ipa-all -fdump-ipa-cgraph @gol
254 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
255 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
256 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
257 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
259 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
260 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
261 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
262 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
263 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
264 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
265 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
266 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
267 -fdump-tree-nrv -fdump-tree-vect @gol
268 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
269 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
270 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
271 -feliminate-unused-debug-symbols -fmem-report -fprofile-arcs -ftree-based-profiling @gol
272 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
273 -ftest-coverage -ftime-report -fvar-tracking @gol
274 -g -g@var{level} -gcoff -gdwarf-2 @gol
275 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
276 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
277 -print-multi-directory -print-multi-lib @gol
278 -print-prog-name=@var{program} -print-search-dirs -Q @gol
281 @item Optimization Options
282 @xref{Optimize Options,,Options that Control Optimization}.
283 @gccoptlist{-falign-functions=@var{n} -falign-jumps=@var{n} @gol
284 -falign-labels=@var{n} -falign-loops=@var{n} @gol
285 -fbounds-check -fmudflap -fmudflapth -fmudflapir @gol
286 -fbranch-probabilities -fprofile-values -fvpt -fbranch-target-load-optimize @gol
287 -fbranch-target-load-optimize2 -fbtr-bb-exclusive @gol
288 -fcaller-saves -fcprop-registers @gol
289 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections @gol
290 -fdelayed-branch -fdelete-null-pointer-checks @gol
291 -fexpensive-optimizations -ffast-math -ffloat-store @gol
292 -fforce-addr -fforce-mem -ffunction-sections @gol
293 -fgcse -fgcse-lm -fgcse-sm -fgcse-las -fgcse-after-reload @gol
294 -floop-optimize -fcrossjumping -fif-conversion -fif-conversion2 @gol
295 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
296 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
297 -fmodulo-sched -fno-branch-count-reg @gol
298 -fno-default-inline -fno-defer-pop -floop-optimize2 -fmove-loop-invariants @gol
299 -fno-function-cse -fno-guess-branch-probability @gol
300 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
301 -funsafe-math-optimizations -ffinite-math-only @gol
302 -fno-trapping-math -fno-zero-initialized-in-bss @gol
303 -fomit-frame-pointer -foptimize-register-move @gol
304 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
305 -fprofile-generate -fprofile-use @gol
306 -fregmove -frename-registers @gol
307 -freorder-blocks -freorder-blocks-and-partition -freorder-functions @gol
308 -frerun-cse-after-loop -frerun-loop-opt @gol
309 -frounding-math -fschedule-insns -fschedule-insns2 @gol
310 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
311 -fsched-spec-load-dangerous @gol
312 -fsched-stalled-insns=@var{n} -sched-stalled-insns-dep=@var{n} @gol
313 -fsched2-use-superblocks @gol
314 -fsched2-use-traces -freschedule-modulo-scheduled-loops @gol
315 -fsignaling-nans -fsingle-precision-constant -fspeculative-prefetching @gol
316 -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps @gol
317 -funroll-all-loops -funroll-loops -fpeel-loops @gol
318 -fsplit-ivs-in-unroller -funswitch-loops @gol
319 -fvariable-expansion-in-unroller @gol
320 -ftree-pre -ftree-ccp -ftree-dce -ftree-loop-optimize @gol
321 -ftree-loop-linear -ftree-loop-im -ftree-loop-ivcanon -fivopts @gol
322 -ftree-dominator-opts -ftree-dse -ftree-copyrename @gol
323 -ftree-ch -ftree-sra -ftree-ter -ftree-lrs -ftree-fre -ftree-vectorize @gol
324 --param @var{name}=@var{value}
325 -O -O0 -O1 -O2 -O3 -Os}
327 @item Preprocessor Options
328 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
329 @gccoptlist{-A@var{question}=@var{answer} @gol
330 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
331 -C -dD -dI -dM -dN @gol
332 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
333 -idirafter @var{dir} @gol
334 -include @var{file} -imacros @var{file} @gol
335 -iprefix @var{file} -iwithprefix @var{dir} @gol
336 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
337 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
338 -P -fworking-directory -remap @gol
339 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
340 -Xpreprocessor @var{option}}
342 @item Assembler Option
343 @xref{Assembler Options,,Passing Options to the Assembler}.
344 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
347 @xref{Link Options,,Options for Linking}.
348 @gccoptlist{@var{object-file-name} -l@var{library} @gol
349 -nostartfiles -nodefaultlibs -nostdlib -pie @gol
350 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
351 -Wl,@var{option} -Xlinker @var{option} @gol
354 @item Directory Options
355 @xref{Directory Options,,Options for Directory Search}.
356 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir} -specs=@var{file} -I-}
359 @c I wrote this xref this way to avoid overfull hbox. -- rms
360 @xref{Target Options}.
361 @gccoptlist{-V @var{version} -b @var{machine}}
363 @item Machine Dependent Options
364 @xref{Submodel Options,,Hardware Models and Configurations}.
365 @c This list is ordered alphanumerically by subsection name.
366 @c Try and put the significant identifier (CPU or system) first,
367 @c so users have a clue at guessing where the ones they want will be.
370 @gccoptlist{-EB -EL @gol
371 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
372 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
375 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
376 -mabi=@var{name} @gol
377 -mapcs-stack-check -mno-apcs-stack-check @gol
378 -mapcs-float -mno-apcs-float @gol
379 -mapcs-reentrant -mno-apcs-reentrant @gol
380 -msched-prolog -mno-sched-prolog @gol
381 -mlittle-endian -mbig-endian -mwords-little-endian @gol
382 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
383 -mthumb-interwork -mno-thumb-interwork @gol
384 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
385 -mstructure-size-boundary=@var{n} @gol
386 -mabort-on-noreturn @gol
387 -mlong-calls -mno-long-calls @gol
388 -msingle-pic-base -mno-single-pic-base @gol
389 -mpic-register=@var{reg} @gol
390 -mnop-fun-dllimport @gol
391 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
392 -mpoke-function-name @gol
394 -mtpcs-frame -mtpcs-leaf-frame @gol
395 -mcaller-super-interworking -mcallee-super-interworking}
398 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
399 -mcall-prologues -mno-tablejump -mtiny-stack -mint8}
402 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
403 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
404 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
405 -mstack-align -mdata-align -mconst-align @gol
406 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
407 -melf -maout -melinux -mlinux -sim -sim2 @gol
408 -mmul-bug-workaround -mno-mul-bug-workaround}
410 @emph{Darwin Options}
411 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
412 -arch_only -bind_at_load -bundle -bundle_loader @gol
413 -client_name -compatibility_version -current_version @gol
415 -dependency-file -dylib_file -dylinker_install_name @gol
416 -dynamic -dynamiclib -exported_symbols_list @gol
417 -filelist -flat_namespace -force_cpusubtype_ALL @gol
418 -force_flat_namespace -headerpad_max_install_names @gol
419 -image_base -init -install_name -keep_private_externs @gol
420 -multi_module -multiply_defined -multiply_defined_unused @gol
421 -noall_load -no_dead_strip_inits_and_terms @gol
422 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
423 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
424 -private_bundle -read_only_relocs -sectalign @gol
425 -sectobjectsymbols -whyload -seg1addr @gol
426 -sectcreate -sectobjectsymbols -sectorder @gol
427 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
428 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
429 -segprot -segs_read_only_addr -segs_read_write_addr @gol
430 -single_module -static -sub_library -sub_umbrella @gol
431 -twolevel_namespace -umbrella -undefined @gol
432 -unexported_symbols_list -weak_reference_mismatches @gol
433 -whatsloaded -F -gused -gfull -mone-byte-bool}
435 @emph{DEC Alpha Options}
436 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
437 -mieee -mieee-with-inexact -mieee-conformant @gol
438 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
439 -mtrap-precision=@var{mode} -mbuild-constants @gol
440 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
441 -mbwx -mmax -mfix -mcix @gol
442 -mfloat-vax -mfloat-ieee @gol
443 -mexplicit-relocs -msmall-data -mlarge-data @gol
444 -msmall-text -mlarge-text @gol
445 -mmemory-latency=@var{time}}
447 @emph{DEC Alpha/VMS Options}
448 @gccoptlist{-mvms-return-codes}
451 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
452 -mhard-float -msoft-float @gol
453 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
454 -mdouble -mno-double @gol
455 -mmedia -mno-media -mmuladd -mno-muladd @gol
456 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
457 -mlinked-fp -mlong-calls -malign-labels @gol
458 -mlibrary-pic -macc-4 -macc-8 @gol
459 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
460 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
461 -mvliw-branch -mno-vliw-branch @gol
462 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
463 -mno-nested-cond-exec -mtomcat-stats @gol
467 @emph{H8/300 Options}
468 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
471 @gccoptlist{-march=@var{architecture-type} @gol
472 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
473 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
474 -mfixed-range=@var{register-range} @gol
475 -mjump-in-delay -mlinker-opt -mlong-calls @gol
476 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
477 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
478 -mno-jump-in-delay -mno-long-load-store @gol
479 -mno-portable-runtime -mno-soft-float @gol
480 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
481 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
482 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
483 -munix=@var{unix-std} -nolibdld -static -threads}
485 @emph{i386 and x86-64 Options}
486 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
487 -mfpmath=@var{unit} @gol
488 -masm=@var{dialect} -mno-fancy-math-387 @gol
489 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
490 -mno-wide-multiply -mrtd -malign-double @gol
491 -mpreferred-stack-boundary=@var{num} @gol
492 -mmmx -msse -msse2 -msse3 -m3dnow @gol
493 -mthreads -mno-align-stringops -minline-all-stringops @gol
494 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
495 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
496 -mno-red-zone -mno-tls-direct-seg-refs @gol
497 -mcmodel=@var{code-model} @gol
501 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
502 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
503 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
504 -minline-float-divide-max-throughput @gol
505 -minline-int-divide-min-latency @gol
506 -minline-int-divide-max-throughput @gol
507 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
508 -mno-dwarf2-asm -mearly-stop-bits @gol
509 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
510 -mtune=@var{cpu-type} -mt -pthread -milp32 -mlp64}
512 @emph{M32R/D Options}
513 @gccoptlist{-m32r2 -m32rx -m32r @gol
515 -malign-loops -mno-align-loops @gol
516 -missue-rate=@var{number} @gol
517 -mbranch-cost=@var{number} @gol
518 -mmodel=@var{code-size-model-type} @gol
519 -msdata=@var{sdata-type} @gol
520 -mno-flush-func -mflush-func=@var{name} @gol
521 -mno-flush-trap -mflush-trap=@var{number} @gol
524 @emph{M680x0 Options}
525 @gccoptlist{-m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
526 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
527 -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
528 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
529 -mshared-library-id=n -mid-shared-library -mno-id-shared-library}
531 @emph{M68hc1x Options}
532 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
533 -mauto-incdec -minmax -mlong-calls -mshort @gol
534 -msoft-reg-count=@var{count}}
537 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
538 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
539 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
540 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
541 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
544 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
545 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
546 -mips16 -mno-mips16 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
547 -mxgot -mno-xgot -mgp32 -mgp64 -mfp32 -mfp64 @gol
548 -mhard-float -msoft-float -msingle-float -mdouble-float @gol
549 -mpaired-single -mips3d @gol
550 -mint64 -mlong64 -mlong32 @gol
551 -G@var{num} -membedded-data -mno-embedded-data @gol
552 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
553 -msplit-addresses -mno-split-addresses @gol
554 -mexplicit-relocs -mno-explicit-relocs @gol
555 -mcheck-zero-division -mno-check-zero-division @gol
556 -mdivide-traps -mdivide-breaks @gol
557 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
558 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
559 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
560 -mfix-vr4120 -mno-fix-vr4120 -mfix-sb1 -mno-fix-sb1 @gol
561 -mflush-func=@var{func} -mno-flush-func @gol
562 -mbranch-likely -mno-branch-likely @gol
563 -mfp-exceptions -mno-fp-exceptions @gol
564 -mvr4130-align -mno-vr4130-align}
567 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
568 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
569 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
570 -mno-base-addresses -msingle-exit -mno-single-exit}
572 @emph{MN10300 Options}
573 @gccoptlist{-mmult-bug -mno-mult-bug @gol
574 -mam33 -mno-am33 @gol
575 -mam33-2 -mno-am33-2 @gol
579 @gccoptlist{-m32032 -m32332 -m32532 -m32081 -m32381 @gol
580 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
581 -mregparam -mnoregparam -msb -mnosb @gol
582 -mbitfield -mnobitfield -mhimem -mnohimem}
584 @emph{PDP-11 Options}
585 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
586 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
587 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
588 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
589 -mbranch-expensive -mbranch-cheap @gol
590 -msplit -mno-split -munix-asm -mdec-asm}
592 @emph{PowerPC Options}
593 See RS/6000 and PowerPC Options.
595 @emph{RS/6000 and PowerPC Options}
596 @gccoptlist{-mcpu=@var{cpu-type} @gol
597 -mtune=@var{cpu-type} @gol
598 -mpower -mno-power -mpower2 -mno-power2 @gol
599 -mpowerpc -mpowerpc64 -mno-powerpc @gol
600 -maltivec -mno-altivec @gol
601 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
602 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
603 -mnew-mnemonics -mold-mnemonics @gol
604 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
605 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
606 -malign-power -malign-natural @gol
607 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
608 -mstring -mno-string -mupdate -mno-update @gol
609 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
610 -mstrict-align -mno-strict-align -mrelocatable @gol
611 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
612 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
613 -mdynamic-no-pic @gol
614 -mprioritize-restricted-insns=@var{priority} @gol
615 -msched-costly-dep=@var{dependence_type} @gol
616 -minsert-sched-nops=@var{scheme} @gol
617 -mcall-sysv -mcall-netbsd @gol
618 -maix-struct-return -msvr4-struct-return @gol
619 -mabi=altivec -mabi=no-altivec @gol
620 -mabi=spe -mabi=no-spe @gol
621 -misel=yes -misel=no @gol
622 -mspe=yes -mspe=no @gol
623 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
624 -mprototype -mno-prototype @gol
625 -msim -mmvme -mads -myellowknife -memb -msdata @gol
626 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
628 @emph{S/390 and zSeries Options}
629 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
630 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
631 -mpacked-stack -mno-packed-stack @gol
632 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
633 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
634 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
635 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
638 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
639 -m4-nofpu -m4-single-only -m4-single -m4 @gol
640 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
641 -m5-64media -m5-64media-nofpu @gol
642 -m5-32media -m5-32media-nofpu @gol
643 -m5-compact -m5-compact-nofpu @gol
644 -mb -ml -mdalign -mrelax @gol
645 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
646 -mieee -misize -mpadstruct -mspace @gol
647 -mprefergot -musermode}
650 @gccoptlist{-mcpu=@var{cpu-type} @gol
651 -mtune=@var{cpu-type} @gol
652 -mcmodel=@var{code-model} @gol
653 -m32 -m64 -mapp-regs -mno-app-regs @gol
654 -mfaster-structs -mno-faster-structs @gol
655 -mfpu -mno-fpu -mhard-float -msoft-float @gol
656 -mhard-quad-float -msoft-quad-float @gol
657 -mimpure-text -mno-impure-text -mlittle-endian @gol
658 -mstack-bias -mno-stack-bias @gol
659 -munaligned-doubles -mno-unaligned-doubles @gol
660 -mv8plus -mno-v8plus -mvis -mno-vis
663 @emph{System V Options}
664 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
666 @emph{TMS320C3x/C4x Options}
667 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
668 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
669 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
670 -mparallel-insns -mparallel-mpy -mpreserve-float}
673 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
674 -mprolog-function -mno-prolog-function -mspace @gol
675 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
676 -mapp-regs -mno-app-regs @gol
677 -mdisable-callt -mno-disable-callt @gol
683 @gccoptlist{-mg -mgnu -munix}
685 @emph{x86-64 Options}
686 See i386 and x86-64 Options.
688 @emph{Xstormy16 Options}
691 @emph{Xtensa Options}
692 @gccoptlist{-mconst16 -mno-const16 @gol
693 -mfused-madd -mno-fused-madd @gol
694 -mtext-section-literals -mno-text-section-literals @gol
695 -mtarget-align -mno-target-align @gol
696 -mlongcalls -mno-longcalls}
698 @emph{zSeries Options}
699 See S/390 and zSeries Options.
701 @item Code Generation Options
702 @xref{Code Gen Options,,Options for Code Generation Conventions}.
703 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
704 -ffixed-@var{reg} -fexceptions @gol
705 -fnon-call-exceptions -funwind-tables @gol
706 -fasynchronous-unwind-tables @gol
707 -finhibit-size-directive -finstrument-functions @gol
708 -fno-common -fno-ident @gol
709 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
710 -freg-struct-return -fshared-data -fshort-enums @gol
711 -fshort-double -fshort-wchar @gol
712 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
713 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
714 -fargument-alias -fargument-noalias @gol
715 -fargument-noalias-global -fleading-underscore @gol
716 -ftls-model=@var{model} @gol
717 -ftrapv -fwrapv -fbounds-check @gol
722 * Overall Options:: Controlling the kind of output:
723 an executable, object files, assembler files,
724 or preprocessed source.
725 * C Dialect Options:: Controlling the variant of C language compiled.
726 * C++ Dialect Options:: Variations on C++.
727 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
729 * Language Independent Options:: Controlling how diagnostics should be
731 * Warning Options:: How picky should the compiler be?
732 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
733 * Optimize Options:: How much optimization?
734 * Preprocessor Options:: Controlling header files and macro definitions.
735 Also, getting dependency information for Make.
736 * Assembler Options:: Passing options to the assembler.
737 * Link Options:: Specifying libraries and so on.
738 * Directory Options:: Where to find header files and libraries.
739 Where to find the compiler executable files.
740 * Spec Files:: How to pass switches to sub-processes.
741 * Target Options:: Running a cross-compiler, or an old version of GCC.
744 @node Overall Options
745 @section Options Controlling the Kind of Output
747 Compilation can involve up to four stages: preprocessing, compilation
748 proper, assembly and linking, always in that order. GCC is capable of
749 preprocessing and compiling several files either into several
750 assembler input files, or into one assembler input file; then each
751 assembler input file produces an object file, and linking combines all
752 the object files (those newly compiled, and those specified as input)
753 into an executable file.
755 @cindex file name suffix
756 For any given input file, the file name suffix determines what kind of
761 C source code which must be preprocessed.
764 C source code which should not be preprocessed.
767 C++ source code which should not be preprocessed.
770 Objective-C source code. Note that you must link with the @file{libobjc}
771 library to make an Objective-C program work.
774 Objective-C source code which should not be preprocessed.
778 Objective-C++ source code. Note that you must link with the @file{libobjc}
779 library to make an Objective-C++ program work. Note that @samp{.M} refers
780 to a literal capital M@.
783 Objective-C++ source code which should not be preprocessed.
786 C, C++, Objective-C or Objective-C++ header file to be turned into a
791 @itemx @var{file}.cxx
792 @itemx @var{file}.cpp
793 @itemx @var{file}.CPP
794 @itemx @var{file}.c++
796 C++ source code which must be preprocessed. Note that in @samp{.cxx},
797 the last two letters must both be literally @samp{x}. Likewise,
798 @samp{.C} refers to a literal capital C@.
802 C++ header file to be turned into a precompiled header.
805 @itemx @var{file}.for
806 @itemx @var{file}.FOR
807 Fortran source code which should not be preprocessed.
810 @itemx @var{file}.fpp
811 @itemx @var{file}.FPP
812 Fortran source code which must be preprocessed (with the traditional
816 Fortran source code which must be preprocessed with a RATFOR
817 preprocessor (not included with GCC)@.
820 @itemx @var{file}.f95
821 Fortran 90/95 source code which should not be preprocessed.
823 @c FIXME: Descriptions of Java file types.
830 Ada source code file which contains a library unit declaration (a
831 declaration of a package, subprogram, or generic, or a generic
832 instantiation), or a library unit renaming declaration (a package,
833 generic, or subprogram renaming declaration). Such files are also
836 @itemx @var{file}.adb
837 Ada source code file containing a library unit body (a subprogram or
838 package body). Such files are also called @dfn{bodies}.
840 @c GCC also knows about some suffixes for languages not yet included:
849 Assembler code which must be preprocessed.
852 An object file to be fed straight into linking.
853 Any file name with no recognized suffix is treated this way.
857 You can specify the input language explicitly with the @option{-x} option:
860 @item -x @var{language}
861 Specify explicitly the @var{language} for the following input files
862 (rather than letting the compiler choose a default based on the file
863 name suffix). This option applies to all following input files until
864 the next @option{-x} option. Possible values for @var{language} are:
866 c c-header c-cpp-output
867 c++ c++-header c++-cpp-output
868 objective-c objective-c-header objective-c-cpp-output
869 objective-c++ objective-c++-header objective-c++-cpp-output
870 assembler assembler-with-cpp
872 f77 f77-cpp-input ratfor
879 Turn off any specification of a language, so that subsequent files are
880 handled according to their file name suffixes (as they are if @option{-x}
881 has not been used at all).
883 @item -pass-exit-codes
884 @opindex pass-exit-codes
885 Normally the @command{gcc} program will exit with the code of 1 if any
886 phase of the compiler returns a non-success return code. If you specify
887 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
888 numerically highest error produced by any phase that returned an error
892 If you only want some of the stages of compilation, you can use
893 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
894 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
895 @command{gcc} is to stop. Note that some combinations (for example,
896 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
901 Compile or assemble the source files, but do not link. The linking
902 stage simply is not done. The ultimate output is in the form of an
903 object file for each source file.
905 By default, the object file name for a source file is made by replacing
906 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
908 Unrecognized input files, not requiring compilation or assembly, are
913 Stop after the stage of compilation proper; do not assemble. The output
914 is in the form of an assembler code file for each non-assembler input
917 By default, the assembler file name for a source file is made by
918 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
920 Input files that don't require compilation are ignored.
924 Stop after the preprocessing stage; do not run the compiler proper. The
925 output is in the form of preprocessed source code, which is sent to the
928 Input files which don't require preprocessing are ignored.
930 @cindex output file option
933 Place output in file @var{file}. This applies regardless to whatever
934 sort of output is being produced, whether it be an executable file,
935 an object file, an assembler file or preprocessed C code.
937 If @option{-o} is not specified, the default is to put an executable
938 file in @file{a.out}, the object file for
939 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
940 assembler file in @file{@var{source}.s}, a precompiled header file in
941 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
946 Print (on standard error output) the commands executed to run the stages
947 of compilation. Also print the version number of the compiler driver
948 program and of the preprocessor and the compiler proper.
952 Like @option{-v} except the commands are not executed and all command
953 arguments are quoted. This is useful for shell scripts to capture the
954 driver-generated command lines.
958 Use pipes rather than temporary files for communication between the
959 various stages of compilation. This fails to work on some systems where
960 the assembler is unable to read from a pipe; but the GNU assembler has
965 If you are compiling multiple source files, this option tells the driver
966 to pass all the source files to the compiler at once (for those
967 languages for which the compiler can handle this). This will allow
968 intermodule analysis (IMA) to be performed by the compiler. Currently the only
969 language for which this is supported is C@. If you pass source files for
970 multiple languages to the driver, using this option, the driver will invoke
971 the compiler(s) that support IMA once each, passing each compiler all the
972 source files appropriate for it. For those languages that do not support
973 IMA this option will be ignored, and the compiler will be invoked once for
974 each source file in that language. If you use this option in conjunction
975 with @option{-save-temps}, the compiler will generate multiple
977 (one for each source file), but only one (combined) @file{.o} or
982 Print (on the standard output) a description of the command line options
983 understood by @command{gcc}. If the @option{-v} option is also specified
984 then @option{--help} will also be passed on to the various processes
985 invoked by @command{gcc}, so that they can display the command line options
986 they accept. If the @option{-Wextra} option is also specified then command
987 line options which have no documentation associated with them will also
992 Print (on the standard output) a description of target specific command
993 line options for each tool.
997 Display the version number and copyrights of the invoked GCC@.
1001 @section Compiling C++ Programs
1003 @cindex suffixes for C++ source
1004 @cindex C++ source file suffixes
1005 C++ source files conventionally use one of the suffixes @samp{.C},
1006 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1007 @samp{.cxx}; C++ header files often use @samp{.hh} or @samp{.H}; and
1008 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1009 files with these names and compiles them as C++ programs even if you
1010 call the compiler the same way as for compiling C programs (usually
1011 with the name @command{gcc}).
1015 However, C++ programs often require class libraries as well as a
1016 compiler that understands the C++ language---and under some
1017 circumstances, you might want to compile programs or header files from
1018 standard input, or otherwise without a suffix that flags them as C++
1019 programs. You might also like to precompile a C header file with a
1020 @samp{.h} extension to be used in C++ compilations. @command{g++} is a
1021 program that calls GCC with the default language set to C++, and
1022 automatically specifies linking against the C++ library. On many
1023 systems, @command{g++} is also installed with the name @command{c++}.
1025 @cindex invoking @command{g++}
1026 When you compile C++ programs, you may specify many of the same
1027 command-line options that you use for compiling programs in any
1028 language; or command-line options meaningful for C and related
1029 languages; or options that are meaningful only for C++ programs.
1030 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1031 explanations of options for languages related to C@.
1032 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1033 explanations of options that are meaningful only for C++ programs.
1035 @node C Dialect Options
1036 @section Options Controlling C Dialect
1037 @cindex dialect options
1038 @cindex language dialect options
1039 @cindex options, dialect
1041 The following options control the dialect of C (or languages derived
1042 from C, such as C++, Objective-C and Objective-C++) that the compiler
1046 @cindex ANSI support
1050 In C mode, support all ISO C90 programs. In C++ mode,
1051 remove GNU extensions that conflict with ISO C++.
1053 This turns off certain features of GCC that are incompatible with ISO
1054 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1055 such as the @code{asm} and @code{typeof} keywords, and
1056 predefined macros such as @code{unix} and @code{vax} that identify the
1057 type of system you are using. It also enables the undesirable and
1058 rarely used ISO trigraph feature. For the C compiler,
1059 it disables recognition of C++ style @samp{//} comments as well as
1060 the @code{inline} keyword.
1062 The alternate keywords @code{__asm__}, @code{__extension__},
1063 @code{__inline__} and @code{__typeof__} continue to work despite
1064 @option{-ansi}. You would not want to use them in an ISO C program, of
1065 course, but it is useful to put them in header files that might be included
1066 in compilations done with @option{-ansi}. Alternate predefined macros
1067 such as @code{__unix__} and @code{__vax__} are also available, with or
1068 without @option{-ansi}.
1070 The @option{-ansi} option does not cause non-ISO programs to be
1071 rejected gratuitously. For that, @option{-pedantic} is required in
1072 addition to @option{-ansi}. @xref{Warning Options}.
1074 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1075 option is used. Some header files may notice this macro and refrain
1076 from declaring certain functions or defining certain macros that the
1077 ISO standard doesn't call for; this is to avoid interfering with any
1078 programs that might use these names for other things.
1080 Functions which would normally be built in but do not have semantics
1081 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1082 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1083 built-in functions provided by GCC}, for details of the functions
1088 Determine the language standard. This option is currently only
1089 supported when compiling C or C++. A value for this option must be
1090 provided; possible values are
1095 ISO C90 (same as @option{-ansi}).
1097 @item iso9899:199409
1098 ISO C90 as modified in amendment 1.
1104 ISO C99. Note that this standard is not yet fully supported; see
1105 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1106 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1109 Default, ISO C90 plus GNU extensions (including some C99 features).
1113 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1114 this will become the default. The name @samp{gnu9x} is deprecated.
1117 The 1998 ISO C++ standard plus amendments.
1120 The same as @option{-std=c++98} plus GNU extensions. This is the
1121 default for C++ code.
1124 Even when this option is not specified, you can still use some of the
1125 features of newer standards in so far as they do not conflict with
1126 previous C standards. For example, you may use @code{__restrict__} even
1127 when @option{-std=c99} is not specified.
1129 The @option{-std} options specifying some version of ISO C have the same
1130 effects as @option{-ansi}, except that features that were not in ISO C90
1131 but are in the specified version (for example, @samp{//} comments and
1132 the @code{inline} keyword in ISO C99) are not disabled.
1134 @xref{Standards,,Language Standards Supported by GCC}, for details of
1135 these standard versions.
1137 @item -aux-info @var{filename}
1139 Output to the given filename prototyped declarations for all functions
1140 declared and/or defined in a translation unit, including those in header
1141 files. This option is silently ignored in any language other than C@.
1143 Besides declarations, the file indicates, in comments, the origin of
1144 each declaration (source file and line), whether the declaration was
1145 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1146 @samp{O} for old, respectively, in the first character after the line
1147 number and the colon), and whether it came from a declaration or a
1148 definition (@samp{C} or @samp{F}, respectively, in the following
1149 character). In the case of function definitions, a K&R-style list of
1150 arguments followed by their declarations is also provided, inside
1151 comments, after the declaration.
1155 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1156 keyword, so that code can use these words as identifiers. You can use
1157 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1158 instead. @option{-ansi} implies @option{-fno-asm}.
1160 In C++, this switch only affects the @code{typeof} keyword, since
1161 @code{asm} and @code{inline} are standard keywords. You may want to
1162 use the @option{-fno-gnu-keywords} flag instead, which has the same
1163 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1164 switch only affects the @code{asm} and @code{typeof} keywords, since
1165 @code{inline} is a standard keyword in ISO C99.
1168 @itemx -fno-builtin-@var{function}
1169 @opindex fno-builtin
1170 @cindex built-in functions
1171 Don't recognize built-in functions that do not begin with
1172 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1173 functions provided by GCC}, for details of the functions affected,
1174 including those which are not built-in functions when @option{-ansi} or
1175 @option{-std} options for strict ISO C conformance are used because they
1176 do not have an ISO standard meaning.
1178 GCC normally generates special code to handle certain built-in functions
1179 more efficiently; for instance, calls to @code{alloca} may become single
1180 instructions that adjust the stack directly, and calls to @code{memcpy}
1181 may become inline copy loops. The resulting code is often both smaller
1182 and faster, but since the function calls no longer appear as such, you
1183 cannot set a breakpoint on those calls, nor can you change the behavior
1184 of the functions by linking with a different library. In addition,
1185 when a function is recognized as a built-in function, GCC may use
1186 information about that function to warn about problems with calls to
1187 that function, or to generate more efficient code, even if the
1188 resulting code still contains calls to that function. For example,
1189 warnings are given with @option{-Wformat} for bad calls to
1190 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1191 known not to modify global memory.
1193 With the @option{-fno-builtin-@var{function}} option
1194 only the built-in function @var{function} is
1195 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1196 function is named this is not built-in in this version of GCC, this
1197 option is ignored. There is no corresponding
1198 @option{-fbuiltin-@var{function}} option; if you wish to enable
1199 built-in functions selectively when using @option{-fno-builtin} or
1200 @option{-ffreestanding}, you may define macros such as:
1203 #define abs(n) __builtin_abs ((n))
1204 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1209 @cindex hosted environment
1211 Assert that compilation takes place in a hosted environment. This implies
1212 @option{-fbuiltin}. A hosted environment is one in which the
1213 entire standard library is available, and in which @code{main} has a return
1214 type of @code{int}. Examples are nearly everything except a kernel.
1215 This is equivalent to @option{-fno-freestanding}.
1217 @item -ffreestanding
1218 @opindex ffreestanding
1219 @cindex hosted environment
1221 Assert that compilation takes place in a freestanding environment. This
1222 implies @option{-fno-builtin}. A freestanding environment
1223 is one in which the standard library may not exist, and program startup may
1224 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1225 This is equivalent to @option{-fno-hosted}.
1227 @xref{Standards,,Language Standards Supported by GCC}, for details of
1228 freestanding and hosted environments.
1230 @item -fms-extensions
1231 @opindex fms-extensions
1232 Accept some non-standard constructs used in Microsoft header files.
1234 Some cases of unnamed fields in structures and unions are only
1235 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1236 fields within structs/unions}, for details.
1240 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1241 options for strict ISO C conformance) implies @option{-trigraphs}.
1243 @item -no-integrated-cpp
1244 @opindex no-integrated-cpp
1245 Performs a compilation in two passes: preprocessing and compiling. This
1246 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1247 @option{-B} option. The user supplied compilation step can then add in
1248 an additional preprocessing step after normal preprocessing but before
1249 compiling. The default is to use the integrated cpp (internal cpp)
1251 The semantics of this option will change if "cc1", "cc1plus", and
1252 "cc1obj" are merged.
1254 @cindex traditional C language
1255 @cindex C language, traditional
1257 @itemx -traditional-cpp
1258 @opindex traditional-cpp
1259 @opindex traditional
1260 Formerly, these options caused GCC to attempt to emulate a pre-standard
1261 C compiler. They are now only supported with the @option{-E} switch.
1262 The preprocessor continues to support a pre-standard mode. See the GNU
1263 CPP manual for details.
1265 @item -fcond-mismatch
1266 @opindex fcond-mismatch
1267 Allow conditional expressions with mismatched types in the second and
1268 third arguments. The value of such an expression is void. This option
1269 is not supported for C++.
1271 @item -funsigned-char
1272 @opindex funsigned-char
1273 Let the type @code{char} be unsigned, like @code{unsigned char}.
1275 Each kind of machine has a default for what @code{char} should
1276 be. It is either like @code{unsigned char} by default or like
1277 @code{signed char} by default.
1279 Ideally, a portable program should always use @code{signed char} or
1280 @code{unsigned char} when it depends on the signedness of an object.
1281 But many programs have been written to use plain @code{char} and
1282 expect it to be signed, or expect it to be unsigned, depending on the
1283 machines they were written for. This option, and its inverse, let you
1284 make such a program work with the opposite default.
1286 The type @code{char} is always a distinct type from each of
1287 @code{signed char} or @code{unsigned char}, even though its behavior
1288 is always just like one of those two.
1291 @opindex fsigned-char
1292 Let the type @code{char} be signed, like @code{signed char}.
1294 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1295 the negative form of @option{-funsigned-char}. Likewise, the option
1296 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1298 @item -fsigned-bitfields
1299 @itemx -funsigned-bitfields
1300 @itemx -fno-signed-bitfields
1301 @itemx -fno-unsigned-bitfields
1302 @opindex fsigned-bitfields
1303 @opindex funsigned-bitfields
1304 @opindex fno-signed-bitfields
1305 @opindex fno-unsigned-bitfields
1306 These options control whether a bit-field is signed or unsigned, when the
1307 declaration does not use either @code{signed} or @code{unsigned}. By
1308 default, such a bit-field is signed, because this is consistent: the
1309 basic integer types such as @code{int} are signed types.
1312 @node C++ Dialect Options
1313 @section Options Controlling C++ Dialect
1315 @cindex compiler options, C++
1316 @cindex C++ options, command line
1317 @cindex options, C++
1318 This section describes the command-line options that are only meaningful
1319 for C++ programs; but you can also use most of the GNU compiler options
1320 regardless of what language your program is in. For example, you
1321 might compile a file @code{firstClass.C} like this:
1324 g++ -g -frepo -O -c firstClass.C
1328 In this example, only @option{-frepo} is an option meant
1329 only for C++ programs; you can use the other options with any
1330 language supported by GCC@.
1332 Here is a list of options that are @emph{only} for compiling C++ programs:
1336 @item -fabi-version=@var{n}
1337 @opindex fabi-version
1338 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1339 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1340 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1341 the version that conforms most closely to the C++ ABI specification.
1342 Therefore, the ABI obtained using version 0 will change as ABI bugs
1345 The default is version 2.
1347 @item -fno-access-control
1348 @opindex fno-access-control
1349 Turn off all access checking. This switch is mainly useful for working
1350 around bugs in the access control code.
1354 Check that the pointer returned by @code{operator new} is non-null
1355 before attempting to modify the storage allocated. This check is
1356 normally unnecessary because the C++ standard specifies that
1357 @code{operator new} will only return @code{0} if it is declared
1358 @samp{throw()}, in which case the compiler will always check the
1359 return value even without this option. In all other cases, when
1360 @code{operator new} has a non-empty exception specification, memory
1361 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1362 @samp{new (nothrow)}.
1364 @item -fconserve-space
1365 @opindex fconserve-space
1366 Put uninitialized or runtime-initialized global variables into the
1367 common segment, as C does. This saves space in the executable at the
1368 cost of not diagnosing duplicate definitions. If you compile with this
1369 flag and your program mysteriously crashes after @code{main()} has
1370 completed, you may have an object that is being destroyed twice because
1371 two definitions were merged.
1373 This option is no longer useful on most targets, now that support has
1374 been added for putting variables into BSS without making them common.
1376 @item -fno-const-strings
1377 @opindex fno-const-strings
1378 Give string constants type @code{char *} instead of type @code{const
1379 char *}. By default, G++ uses type @code{const char *} as required by
1380 the standard. Even if you use @option{-fno-const-strings}, you cannot
1381 actually modify the value of a string constant.
1383 This option might be removed in a future release of G++. For maximum
1384 portability, you should structure your code so that it works with
1385 string constants that have type @code{const char *}.
1387 @item -fno-elide-constructors
1388 @opindex fno-elide-constructors
1389 The C++ standard allows an implementation to omit creating a temporary
1390 which is only used to initialize another object of the same type.
1391 Specifying this option disables that optimization, and forces G++ to
1392 call the copy constructor in all cases.
1394 @item -fno-enforce-eh-specs
1395 @opindex fno-enforce-eh-specs
1396 Don't check for violation of exception specifications at runtime. This
1397 option violates the C++ standard, but may be useful for reducing code
1398 size in production builds, much like defining @samp{NDEBUG}. The compiler
1399 will still optimize based on the exception specifications.
1402 @itemx -fno-for-scope
1404 @opindex fno-for-scope
1405 If @option{-ffor-scope} is specified, the scope of variables declared in
1406 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1407 as specified by the C++ standard.
1408 If @option{-fno-for-scope} is specified, the scope of variables declared in
1409 a @i{for-init-statement} extends to the end of the enclosing scope,
1410 as was the case in old versions of G++, and other (traditional)
1411 implementations of C++.
1413 The default if neither flag is given to follow the standard,
1414 but to allow and give a warning for old-style code that would
1415 otherwise be invalid, or have different behavior.
1417 @item -fno-gnu-keywords
1418 @opindex fno-gnu-keywords
1419 Do not recognize @code{typeof} as a keyword, so that code can use this
1420 word as an identifier. You can use the keyword @code{__typeof__} instead.
1421 @option{-ansi} implies @option{-fno-gnu-keywords}.
1423 @item -fno-implicit-templates
1424 @opindex fno-implicit-templates
1425 Never emit code for non-inline templates which are instantiated
1426 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1427 @xref{Template Instantiation}, for more information.
1429 @item -fno-implicit-inline-templates
1430 @opindex fno-implicit-inline-templates
1431 Don't emit code for implicit instantiations of inline templates, either.
1432 The default is to handle inlines differently so that compiles with and
1433 without optimization will need the same set of explicit instantiations.
1435 @item -fno-implement-inlines
1436 @opindex fno-implement-inlines
1437 To save space, do not emit out-of-line copies of inline functions
1438 controlled by @samp{#pragma implementation}. This will cause linker
1439 errors if these functions are not inlined everywhere they are called.
1441 @item -fms-extensions
1442 @opindex fms-extensions
1443 Disable pedantic warnings about constructs used in MFC, such as implicit
1444 int and getting a pointer to member function via non-standard syntax.
1446 @item -fno-nonansi-builtins
1447 @opindex fno-nonansi-builtins
1448 Disable built-in declarations of functions that are not mandated by
1449 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1450 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1452 @item -fno-operator-names
1453 @opindex fno-operator-names
1454 Do not treat the operator name keywords @code{and}, @code{bitand},
1455 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1456 synonyms as keywords.
1458 @item -fno-optional-diags
1459 @opindex fno-optional-diags
1460 Disable diagnostics that the standard says a compiler does not need to
1461 issue. Currently, the only such diagnostic issued by G++ is the one for
1462 a name having multiple meanings within a class.
1465 @opindex fpermissive
1466 Downgrade some diagnostics about nonconformant code from errors to
1467 warnings. Thus, using @option{-fpermissive} will allow some
1468 nonconforming code to compile.
1472 Enable automatic template instantiation at link time. This option also
1473 implies @option{-fno-implicit-templates}. @xref{Template
1474 Instantiation}, for more information.
1478 Disable generation of information about every class with virtual
1479 functions for use by the C++ runtime type identification features
1480 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1481 of the language, you can save some space by using this flag. Note that
1482 exception handling uses the same information, but it will generate it as
1487 Emit statistics about front-end processing at the end of the compilation.
1488 This information is generally only useful to the G++ development team.
1490 @item -ftemplate-depth-@var{n}
1491 @opindex ftemplate-depth
1492 Set the maximum instantiation depth for template classes to @var{n}.
1493 A limit on the template instantiation depth is needed to detect
1494 endless recursions during template class instantiation. ANSI/ISO C++
1495 conforming programs must not rely on a maximum depth greater than 17.
1497 @item -fno-threadsafe-statics
1498 @opindex fno-threadsafe-statics
1499 Do not emit the extra code to use the routines specified in the C++
1500 ABI for thread-safe initialization of local statics. You can use this
1501 option to reduce code size slightly in code that doesn't need to be
1504 @item -fuse-cxa-atexit
1505 @opindex fuse-cxa-atexit
1506 Register destructors for objects with static storage duration with the
1507 @code{__cxa_atexit} function rather than the @code{atexit} function.
1508 This option is required for fully standards-compliant handling of static
1509 destructors, but will only work if your C library supports
1510 @code{__cxa_atexit}.
1512 @item -fvisibility-inlines-hidden
1513 @opindex fvisibility-inlines-hidden
1514 Causes all inlined methods to be marked with
1515 @code{__attribute__ ((visibility ("hidden")))} so that they do not
1516 appear in the export table of a DSO and do not require a PLT indirection
1517 when used within the DSO@. Enabling this option can have a dramatic effect
1518 on load and link times of a DSO as it massively reduces the size of the
1519 dynamic export table when the library makes heavy use of templates. While
1520 it can cause bloating through duplication of code within each DSO where
1521 it is used, often the wastage is less than the considerable space occupied
1522 by a long symbol name in the export table which is typical when using
1523 templates and namespaces. For even more savings, combine with the
1524 @option{-fvisibility=hidden} switch.
1528 Do not use weak symbol support, even if it is provided by the linker.
1529 By default, G++ will use weak symbols if they are available. This
1530 option exists only for testing, and should not be used by end-users;
1531 it will result in inferior code and has no benefits. This option may
1532 be removed in a future release of G++.
1536 Do not search for header files in the standard directories specific to
1537 C++, but do still search the other standard directories. (This option
1538 is used when building the C++ library.)
1541 In addition, these optimization, warning, and code generation options
1542 have meanings only for C++ programs:
1545 @item -fno-default-inline
1546 @opindex fno-default-inline
1547 Do not assume @samp{inline} for functions defined inside a class scope.
1548 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1549 functions will have linkage like inline functions; they just won't be
1552 @item -Wabi @r{(C++ only)}
1554 Warn when G++ generates code that is probably not compatible with the
1555 vendor-neutral C++ ABI@. Although an effort has been made to warn about
1556 all such cases, there are probably some cases that are not warned about,
1557 even though G++ is generating incompatible code. There may also be
1558 cases where warnings are emitted even though the code that is generated
1561 You should rewrite your code to avoid these warnings if you are
1562 concerned about the fact that code generated by G++ may not be binary
1563 compatible with code generated by other compilers.
1565 The known incompatibilities at this point include:
1570 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1571 pack data into the same byte as a base class. For example:
1574 struct A @{ virtual void f(); int f1 : 1; @};
1575 struct B : public A @{ int f2 : 1; @};
1579 In this case, G++ will place @code{B::f2} into the same byte
1580 as@code{A::f1}; other compilers will not. You can avoid this problem
1581 by explicitly padding @code{A} so that its size is a multiple of the
1582 byte size on your platform; that will cause G++ and other compilers to
1583 layout @code{B} identically.
1586 Incorrect handling of tail-padding for virtual bases. G++ does not use
1587 tail padding when laying out virtual bases. For example:
1590 struct A @{ virtual void f(); char c1; @};
1591 struct B @{ B(); char c2; @};
1592 struct C : public A, public virtual B @{@};
1596 In this case, G++ will not place @code{B} into the tail-padding for
1597 @code{A}; other compilers will. You can avoid this problem by
1598 explicitly padding @code{A} so that its size is a multiple of its
1599 alignment (ignoring virtual base classes); that will cause G++ and other
1600 compilers to layout @code{C} identically.
1603 Incorrect handling of bit-fields with declared widths greater than that
1604 of their underlying types, when the bit-fields appear in a union. For
1608 union U @{ int i : 4096; @};
1612 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1613 union too small by the number of bits in an @code{int}.
1616 Empty classes can be placed at incorrect offsets. For example:
1626 struct C : public B, public A @{@};
1630 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1631 it should be placed at offset zero. G++ mistakenly believes that the
1632 @code{A} data member of @code{B} is already at offset zero.
1635 Names of template functions whose types involve @code{typename} or
1636 template template parameters can be mangled incorrectly.
1639 template <typename Q>
1640 void f(typename Q::X) @{@}
1642 template <template <typename> class Q>
1643 void f(typename Q<int>::X) @{@}
1647 Instantiations of these templates may be mangled incorrectly.
1651 @item -Wctor-dtor-privacy @r{(C++ only)}
1652 @opindex Wctor-dtor-privacy
1653 Warn when a class seems unusable because all the constructors or
1654 destructors in that class are private, and it has neither friends nor
1655 public static member functions.
1657 @item -Wnon-virtual-dtor @r{(C++ only)}
1658 @opindex Wnon-virtual-dtor
1659 Warn when a class appears to be polymorphic, thereby requiring a virtual
1660 destructor, yet it declares a non-virtual one.
1661 This warning is enabled by @option{-Wall}.
1663 @item -Wreorder @r{(C++ only)}
1665 @cindex reordering, warning
1666 @cindex warning for reordering of member initializers
1667 Warn when the order of member initializers given in the code does not
1668 match the order in which they must be executed. For instance:
1674 A(): j (0), i (1) @{ @}
1678 The compiler will rearrange the member initializers for @samp{i}
1679 and @samp{j} to match the declaration order of the members, emitting
1680 a warning to that effect. This warning is enabled by @option{-Wall}.
1683 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1686 @item -Weffc++ @r{(C++ only)}
1688 Warn about violations of the following style guidelines from Scott Meyers'
1689 @cite{Effective C++} book:
1693 Item 11: Define a copy constructor and an assignment operator for classes
1694 with dynamically allocated memory.
1697 Item 12: Prefer initialization to assignment in constructors.
1700 Item 14: Make destructors virtual in base classes.
1703 Item 15: Have @code{operator=} return a reference to @code{*this}.
1706 Item 23: Don't try to return a reference when you must return an object.
1710 Also warn about violations of the following style guidelines from
1711 Scott Meyers' @cite{More Effective C++} book:
1715 Item 6: Distinguish between prefix and postfix forms of increment and
1716 decrement operators.
1719 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1723 When selecting this option, be aware that the standard library
1724 headers do not obey all of these guidelines; use @samp{grep -v}
1725 to filter out those warnings.
1727 @item -Wno-deprecated @r{(C++ only)}
1728 @opindex Wno-deprecated
1729 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1731 @item -Wno-non-template-friend @r{(C++ only)}
1732 @opindex Wno-non-template-friend
1733 Disable warnings when non-templatized friend functions are declared
1734 within a template. Since the advent of explicit template specification
1735 support in G++, if the name of the friend is an unqualified-id (i.e.,
1736 @samp{friend foo(int)}), the C++ language specification demands that the
1737 friend declare or define an ordinary, nontemplate function. (Section
1738 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1739 could be interpreted as a particular specialization of a templatized
1740 function. Because this non-conforming behavior is no longer the default
1741 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1742 check existing code for potential trouble spots and is on by default.
1743 This new compiler behavior can be turned off with
1744 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1745 but disables the helpful warning.
1747 @item -Wold-style-cast @r{(C++ only)}
1748 @opindex Wold-style-cast
1749 Warn if an old-style (C-style) cast to a non-void type is used within
1750 a C++ program. The new-style casts (@samp{static_cast},
1751 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1752 unintended effects and much easier to search for.
1754 @item -Woverloaded-virtual @r{(C++ only)}
1755 @opindex Woverloaded-virtual
1756 @cindex overloaded virtual fn, warning
1757 @cindex warning for overloaded virtual fn
1758 Warn when a function declaration hides virtual functions from a
1759 base class. For example, in:
1766 struct B: public A @{
1771 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1779 will fail to compile.
1781 @item -Wno-pmf-conversions @r{(C++ only)}
1782 @opindex Wno-pmf-conversions
1783 Disable the diagnostic for converting a bound pointer to member function
1786 @item -Wsign-promo @r{(C++ only)}
1787 @opindex Wsign-promo
1788 Warn when overload resolution chooses a promotion from unsigned or
1789 enumerated type to a signed type, over a conversion to an unsigned type of
1790 the same size. Previous versions of G++ would try to preserve
1791 unsignedness, but the standard mandates the current behavior.
1793 @item -Wsynth @r{(C++ only)}
1795 @cindex warning for synthesized methods
1796 @cindex synthesized methods, warning
1797 Warn when G++'s synthesis behavior does not match that of cfront. For
1803 A& operator = (int);
1813 In this example, G++ will synthesize a default @samp{A& operator =
1814 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1817 @node Objective-C and Objective-C++ Dialect Options
1818 @section Options Controlling Objective-C and Objective-C++ Dialects
1820 @cindex compiler options, Objective-C and Objective-C++
1821 @cindex Objective-C and Objective-C++ options, command line
1822 @cindex options, Objective-C and Objective-C++
1823 (NOTE: This manual does not describe the Objective-C and Objective-C++
1824 languages themselves. See @xref{Standards,,Language Standards
1825 Supported by GCC}, for references.)
1827 This section describes the command-line options that are only meaningful
1828 for Objective-C and Objective-C++ programs, but you can also use most of
1829 the language-independent GNU compiler options.
1830 For example, you might compile a file @code{some_class.m} like this:
1833 gcc -g -fgnu-runtime -O -c some_class.m
1837 In this example, @option{-fgnu-runtime} is an option meant only for
1838 Objective-C and Objective-C++ programs; you can use the other options with
1839 any language supported by GCC@.
1841 Note that since Objective-C is an extension of the C language, Objective-C
1842 compilations may also use options specific to the C front-end (e.g.,
1843 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
1844 C++-specific options (e.g., @option{-Wabi}).
1846 Here is a list of options that are @emph{only} for compiling Objective-C
1847 and Objective-C++ programs:
1850 @item -fconstant-string-class=@var{class-name}
1851 @opindex fconstant-string-class
1852 Use @var{class-name} as the name of the class to instantiate for each
1853 literal string specified with the syntax @code{@@"@dots{}"}. The default
1854 class name is @code{NXConstantString} if the GNU runtime is being used, and
1855 @code{NSConstantString} if the NeXT runtime is being used (see below). The
1856 @option{-fconstant-cfstrings} option, if also present, will override the
1857 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
1858 to be laid out as constant CoreFoundation strings.
1861 @opindex fgnu-runtime
1862 Generate object code compatible with the standard GNU Objective-C
1863 runtime. This is the default for most types of systems.
1865 @item -fnext-runtime
1866 @opindex fnext-runtime
1867 Generate output compatible with the NeXT runtime. This is the default
1868 for NeXT-based systems, including Darwin and Mac OS X@. The macro
1869 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
1872 @item -fno-nil-receivers
1873 @opindex fno-nil-receivers
1874 Assume that all Objective-C message dispatches (e.g.,
1875 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
1876 is not @code{nil}. This allows for more efficient entry points in the runtime
1877 to be used. Currently, this option is only available in conjunction with
1878 the NeXT runtime on Mac OS X 10.3 and later.
1880 @item -fobjc-exceptions
1881 @opindex fobjc-exceptions
1882 Enable syntactic support for structured exception handling in Objective-C,
1883 similar to what is offered by C++ and Java. Currently, this option is only
1884 available in conjunction with the NeXT runtime on Mac OS X 10.3 and later.
1892 @@catch (AnObjCClass *exc) @{
1899 @@catch (AnotherClass *exc) @{
1902 @@catch (id allOthers) @{
1912 The @code{@@throw} statement may appear anywhere in an Objective-C or
1913 Objective-C++ program; when used inside of a @code{@@catch} block, the
1914 @code{@@throw} may appear without an argument (as shown above), in which case
1915 the object caught by the @code{@@catch} will be rethrown.
1917 Note that only (pointers to) Objective-C objects may be thrown and
1918 caught using this scheme. When an object is thrown, it will be caught
1919 by the nearest @code{@@catch} clause capable of handling objects of that type,
1920 analogously to how @code{catch} blocks work in C++ and Java. A
1921 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
1922 any and all Objective-C exceptions not caught by previous @code{@@catch}
1925 The @code{@@finally} clause, if present, will be executed upon exit from the
1926 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
1927 regardless of whether any exceptions are thrown, caught or rethrown
1928 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
1929 of the @code{finally} clause in Java.
1931 There are several caveats to using the new exception mechanism:
1935 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
1936 idioms provided by the @code{NSException} class, the new
1937 exceptions can only be used on Mac OS X 10.3 (Panther) and later
1938 systems, due to additional functionality needed in the (NeXT) Objective-C
1942 As mentioned above, the new exceptions do not support handling
1943 types other than Objective-C objects. Furthermore, when used from
1944 Objective-C++, the Objective-C exception model does not interoperate with C++
1945 exceptions at this time. This means you cannot @code{@@throw} an exception
1946 from Objective-C and @code{catch} it in C++, or vice versa
1947 (i.e., @code{throw @dots{} @@catch}).
1950 The @option{-fobjc-exceptions} switch also enables the use of synchronization
1951 blocks for thread-safe execution:
1954 @@synchronized (ObjCClass *guard) @{
1959 Upon entering the @code{@@synchronized} block, a thread of execution shall
1960 first check whether a lock has been placed on the corresponding @code{guard}
1961 object by another thread. If it has, the current thread shall wait until
1962 the other thread relinquishes its lock. Once @code{guard} becomes available,
1963 the current thread will place its own lock on it, execute the code contained in
1964 the @code{@@synchronized} block, and finally relinquish the lock (thereby
1965 making @code{guard} available to other threads).
1967 Unlike Java, Objective-C does not allow for entire methods to be marked
1968 @code{@@synchronized}. Note that throwing exceptions out of
1969 @code{@@synchronized} blocks is allowed, and will cause the guarding object
1970 to be unlocked properly.
1972 @item -freplace-objc-classes
1973 @opindex freplace-objc-classes
1974 Emit a special marker instructing @command{ld(1)} not to statically link in
1975 the resulting object file, and allow @command{dyld(1)} to load it in at
1976 run time instead. This is used in conjunction with the Fix-and-Continue
1977 debugging mode, where the object file in question may be recompiled and
1978 dynamically reloaded in the course of program execution, without the need
1979 to restart the program itself. Currently, Fix-and-Continue functionality
1980 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
1985 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
1986 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
1987 compile time) with static class references that get initialized at load time,
1988 which improves run-time performance. Specifying the @option{-fzero-link} flag
1989 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
1990 to be retained. This is useful in Zero-Link debugging mode, since it allows
1991 for individual class implementations to be modified during program execution.
1995 Dump interface declarations for all classes seen in the source file to a
1996 file named @file{@var{sourcename}.decl}.
1999 @opindex Wno-protocol
2000 If a class is declared to implement a protocol, a warning is issued for
2001 every method in the protocol that is not implemented by the class. The
2002 default behavior is to issue a warning for every method not explicitly
2003 implemented in the class, even if a method implementation is inherited
2004 from the superclass. If you use the @option{-Wno-protocol} option, then
2005 methods inherited from the superclass are considered to be implemented,
2006 and no warning is issued for them.
2010 Warn if multiple methods of different types for the same selector are
2011 found during compilation. The check is performed on the list of methods
2012 in the final stage of compilation. Additionally, a check is performed
2013 for each selector appearing in a @code{@@selector(@dots{})}
2014 expression, and a corresponding method for that selector has been found
2015 during compilation. Because these checks scan the method table only at
2016 the end of compilation, these warnings are not produced if the final
2017 stage of compilation is not reached, for example because an error is
2018 found during compilation, or because the @option{-fsyntax-only} option is
2021 @item -Wundeclared-selector
2022 @opindex Wundeclared-selector
2023 Warn if a @code{@@selector(@dots{})} expression referring to an
2024 undeclared selector is found. A selector is considered undeclared if no
2025 method with that name has been declared before the
2026 @code{@@selector(@dots{})} expression, either explicitly in an
2027 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2028 an @code{@@implementation} section. This option always performs its
2029 checks as soon as a @code{@@selector(@dots{})} expression is found,
2030 while @option{-Wselector} only performs its checks in the final stage of
2031 compilation. This also enforces the coding style convention
2032 that methods and selectors must be declared before being used.
2034 @item -print-objc-runtime-info
2035 @opindex print-objc-runtime-info
2036 Generate C header describing the largest structure that is passed by
2041 @node Language Independent Options
2042 @section Options to Control Diagnostic Messages Formatting
2043 @cindex options to control diagnostics formatting
2044 @cindex diagnostic messages
2045 @cindex message formatting
2047 Traditionally, diagnostic messages have been formatted irrespective of
2048 the output device's aspect (e.g.@: its width, @dots{}). The options described
2049 below can be used to control the diagnostic messages formatting
2050 algorithm, e.g.@: how many characters per line, how often source location
2051 information should be reported. Right now, only the C++ front end can
2052 honor these options. However it is expected, in the near future, that
2053 the remaining front ends would be able to digest them correctly.
2056 @item -fmessage-length=@var{n}
2057 @opindex fmessage-length
2058 Try to format error messages so that they fit on lines of about @var{n}
2059 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2060 the front ends supported by GCC@. If @var{n} is zero, then no
2061 line-wrapping will be done; each error message will appear on a single
2064 @opindex fdiagnostics-show-location
2065 @item -fdiagnostics-show-location=once
2066 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2067 reporter to emit @emph{once} source location information; that is, in
2068 case the message is too long to fit on a single physical line and has to
2069 be wrapped, the source location won't be emitted (as prefix) again,
2070 over and over, in subsequent continuation lines. This is the default
2073 @item -fdiagnostics-show-location=every-line
2074 Only meaningful in line-wrapping mode. Instructs the diagnostic
2075 messages reporter to emit the same source location information (as
2076 prefix) for physical lines that result from the process of breaking
2077 a message which is too long to fit on a single line.
2081 @node Warning Options
2082 @section Options to Request or Suppress Warnings
2083 @cindex options to control warnings
2084 @cindex warning messages
2085 @cindex messages, warning
2086 @cindex suppressing warnings
2088 Warnings are diagnostic messages that report constructions which
2089 are not inherently erroneous but which are risky or suggest there
2090 may have been an error.
2092 You can request many specific warnings with options beginning @samp{-W},
2093 for example @option{-Wimplicit} to request warnings on implicit
2094 declarations. Each of these specific warning options also has a
2095 negative form beginning @samp{-Wno-} to turn off warnings;
2096 for example, @option{-Wno-implicit}. This manual lists only one of the
2097 two forms, whichever is not the default.
2099 The following options control the amount and kinds of warnings produced
2100 by GCC; for further, language-specific options also refer to
2101 @ref{C++ Dialect Options} and @ref{Objective-C and Objective-C++ Dialect
2105 @cindex syntax checking
2107 @opindex fsyntax-only
2108 Check the code for syntax errors, but don't do anything beyond that.
2112 Issue all the warnings demanded by strict ISO C and ISO C++;
2113 reject all programs that use forbidden extensions, and some other
2114 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2115 version of the ISO C standard specified by any @option{-std} option used.
2117 Valid ISO C and ISO C++ programs should compile properly with or without
2118 this option (though a rare few will require @option{-ansi} or a
2119 @option{-std} option specifying the required version of ISO C)@. However,
2120 without this option, certain GNU extensions and traditional C and C++
2121 features are supported as well. With this option, they are rejected.
2123 @option{-pedantic} does not cause warning messages for use of the
2124 alternate keywords whose names begin and end with @samp{__}. Pedantic
2125 warnings are also disabled in the expression that follows
2126 @code{__extension__}. However, only system header files should use
2127 these escape routes; application programs should avoid them.
2128 @xref{Alternate Keywords}.
2130 Some users try to use @option{-pedantic} to check programs for strict ISO
2131 C conformance. They soon find that it does not do quite what they want:
2132 it finds some non-ISO practices, but not all---only those for which
2133 ISO C @emph{requires} a diagnostic, and some others for which
2134 diagnostics have been added.
2136 A feature to report any failure to conform to ISO C might be useful in
2137 some instances, but would require considerable additional work and would
2138 be quite different from @option{-pedantic}. We don't have plans to
2139 support such a feature in the near future.
2141 Where the standard specified with @option{-std} represents a GNU
2142 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2143 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2144 extended dialect is based. Warnings from @option{-pedantic} are given
2145 where they are required by the base standard. (It would not make sense
2146 for such warnings to be given only for features not in the specified GNU
2147 C dialect, since by definition the GNU dialects of C include all
2148 features the compiler supports with the given option, and there would be
2149 nothing to warn about.)
2151 @item -pedantic-errors
2152 @opindex pedantic-errors
2153 Like @option{-pedantic}, except that errors are produced rather than
2158 Inhibit all warning messages.
2162 Inhibit warning messages about the use of @samp{#import}.
2164 @item -Wchar-subscripts
2165 @opindex Wchar-subscripts
2166 Warn if an array subscript has type @code{char}. This is a common cause
2167 of error, as programmers often forget that this type is signed on some
2172 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2173 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2175 @item -Wfatal-errors
2176 @opindex Wfatal-errors
2177 This option causes the compiler to abort compilation on the first error
2178 occurred rather than trying to keep going and printing further error
2183 @opindex ffreestanding
2184 @opindex fno-builtin
2185 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2186 the arguments supplied have types appropriate to the format string
2187 specified, and that the conversions specified in the format string make
2188 sense. This includes standard functions, and others specified by format
2189 attributes (@pxref{Function Attributes}), in the @code{printf},
2190 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2191 not in the C standard) families (or other target-specific families).
2192 Which functions are checked without format attributes having been
2193 specified depends on the standard version selected, and such checks of
2194 functions without the attribute specified are disabled by
2195 @option{-ffreestanding} or @option{-fno-builtin}.
2197 The formats are checked against the format features supported by GNU
2198 libc version 2.2. These include all ISO C90 and C99 features, as well
2199 as features from the Single Unix Specification and some BSD and GNU
2200 extensions. Other library implementations may not support all these
2201 features; GCC does not support warning about features that go beyond a
2202 particular library's limitations. However, if @option{-pedantic} is used
2203 with @option{-Wformat}, warnings will be given about format features not
2204 in the selected standard version (but not for @code{strfmon} formats,
2205 since those are not in any version of the C standard). @xref{C Dialect
2206 Options,,Options Controlling C Dialect}.
2208 Since @option{-Wformat} also checks for null format arguments for
2209 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2211 @option{-Wformat} is included in @option{-Wall}. For more control over some
2212 aspects of format checking, the options @option{-Wformat-y2k},
2213 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2214 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2215 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2218 @opindex Wformat-y2k
2219 If @option{-Wformat} is specified, also warn about @code{strftime}
2220 formats which may yield only a two-digit year.
2222 @item -Wno-format-extra-args
2223 @opindex Wno-format-extra-args
2224 If @option{-Wformat} is specified, do not warn about excess arguments to a
2225 @code{printf} or @code{scanf} format function. The C standard specifies
2226 that such arguments are ignored.
2228 Where the unused arguments lie between used arguments that are
2229 specified with @samp{$} operand number specifications, normally
2230 warnings are still given, since the implementation could not know what
2231 type to pass to @code{va_arg} to skip the unused arguments. However,
2232 in the case of @code{scanf} formats, this option will suppress the
2233 warning if the unused arguments are all pointers, since the Single
2234 Unix Specification says that such unused arguments are allowed.
2236 @item -Wno-format-zero-length
2237 @opindex Wno-format-zero-length
2238 If @option{-Wformat} is specified, do not warn about zero-length formats.
2239 The C standard specifies that zero-length formats are allowed.
2241 @item -Wformat-nonliteral
2242 @opindex Wformat-nonliteral
2243 If @option{-Wformat} is specified, also warn if the format string is not a
2244 string literal and so cannot be checked, unless the format function
2245 takes its format arguments as a @code{va_list}.
2247 @item -Wformat-security
2248 @opindex Wformat-security
2249 If @option{-Wformat} is specified, also warn about uses of format
2250 functions that represent possible security problems. At present, this
2251 warns about calls to @code{printf} and @code{scanf} functions where the
2252 format string is not a string literal and there are no format arguments,
2253 as in @code{printf (foo);}. This may be a security hole if the format
2254 string came from untrusted input and contains @samp{%n}. (This is
2255 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2256 in future warnings may be added to @option{-Wformat-security} that are not
2257 included in @option{-Wformat-nonliteral}.)
2261 Enable @option{-Wformat} plus format checks not included in
2262 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2263 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
2267 Warn about passing a null pointer for arguments marked as
2268 requiring a non-null value by the @code{nonnull} function attribute.
2270 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2271 can be disabled with the @option{-Wno-nonnull} option.
2273 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
2275 Warn about uninitialized variables which are initialized with themselves.
2276 Note this option can only be used with the @option{-Wuninitialized} option,
2277 which in turn only works with @option{-O1} and above.
2279 For example, GCC will warn about @code{i} being uninitialized in the
2280 following snippet only when @option{-Winit-self} has been specified:
2291 @item -Wimplicit-int
2292 @opindex Wimplicit-int
2293 Warn when a declaration does not specify a type.
2295 @item -Wimplicit-function-declaration
2296 @itemx -Werror-implicit-function-declaration
2297 @opindex Wimplicit-function-declaration
2298 @opindex Werror-implicit-function-declaration
2299 Give a warning (or error) whenever a function is used before being
2300 declared. The form @option{-Wno-error-implicit-function-declaration}
2305 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2309 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2310 function with external linkage, returning int, taking either zero
2311 arguments, two, or three arguments of appropriate types.
2313 @item -Wmissing-braces
2314 @opindex Wmissing-braces
2315 Warn if an aggregate or union initializer is not fully bracketed. In
2316 the following example, the initializer for @samp{a} is not fully
2317 bracketed, but that for @samp{b} is fully bracketed.
2320 int a[2][2] = @{ 0, 1, 2, 3 @};
2321 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2324 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
2325 @opindex Wmissing-include-dirs
2326 Warn if a user-supplied include directory does not exist.
2329 @opindex Wparentheses
2330 Warn if parentheses are omitted in certain contexts, such
2331 as when there is an assignment in a context where a truth value
2332 is expected, or when operators are nested whose precedence people
2333 often get confused about. Only the warning for an assignment used as
2334 a truth value is supported when compiling C++; the other warnings are
2335 only supported when compiling C@.
2337 Also warn if a comparison like @samp{x<=y<=z} appears; this is
2338 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
2339 interpretation from that of ordinary mathematical notation.
2341 Also warn about constructions where there may be confusion to which
2342 @code{if} statement an @code{else} branch belongs. Here is an example of
2357 In C, every @code{else} branch belongs to the innermost possible @code{if}
2358 statement, which in this example is @code{if (b)}. This is often not
2359 what the programmer expected, as illustrated in the above example by
2360 indentation the programmer chose. When there is the potential for this
2361 confusion, GCC will issue a warning when this flag is specified.
2362 To eliminate the warning, add explicit braces around the innermost
2363 @code{if} statement so there is no way the @code{else} could belong to
2364 the enclosing @code{if}. The resulting code would look like this:
2380 @item -Wsequence-point
2381 @opindex Wsequence-point
2382 Warn about code that may have undefined semantics because of violations
2383 of sequence point rules in the C standard.
2385 The C standard defines the order in which expressions in a C program are
2386 evaluated in terms of @dfn{sequence points}, which represent a partial
2387 ordering between the execution of parts of the program: those executed
2388 before the sequence point, and those executed after it. These occur
2389 after the evaluation of a full expression (one which is not part of a
2390 larger expression), after the evaluation of the first operand of a
2391 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2392 function is called (but after the evaluation of its arguments and the
2393 expression denoting the called function), and in certain other places.
2394 Other than as expressed by the sequence point rules, the order of
2395 evaluation of subexpressions of an expression is not specified. All
2396 these rules describe only a partial order rather than a total order,
2397 since, for example, if two functions are called within one expression
2398 with no sequence point between them, the order in which the functions
2399 are called is not specified. However, the standards committee have
2400 ruled that function calls do not overlap.
2402 It is not specified when between sequence points modifications to the
2403 values of objects take effect. Programs whose behavior depends on this
2404 have undefined behavior; the C standard specifies that ``Between the
2405 previous and next sequence point an object shall have its stored value
2406 modified at most once by the evaluation of an expression. Furthermore,
2407 the prior value shall be read only to determine the value to be
2408 stored.''. If a program breaks these rules, the results on any
2409 particular implementation are entirely unpredictable.
2411 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2412 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2413 diagnosed by this option, and it may give an occasional false positive
2414 result, but in general it has been found fairly effective at detecting
2415 this sort of problem in programs.
2417 The present implementation of this option only works for C programs. A
2418 future implementation may also work for C++ programs.
2420 The C standard is worded confusingly, therefore there is some debate
2421 over the precise meaning of the sequence point rules in subtle cases.
2422 Links to discussions of the problem, including proposed formal
2423 definitions, may be found on the GCC readings page, at
2424 @w{@uref{http://gcc.gnu.org/readings.html}}.
2427 @opindex Wreturn-type
2428 Warn whenever a function is defined with a return-type that defaults to
2429 @code{int}. Also warn about any @code{return} statement with no
2430 return-value in a function whose return-type is not @code{void}.
2432 For C, also warn if the return type of a function has a type qualifier
2433 such as @code{const}. Such a type qualifier has no effect, since the
2434 value returned by a function is not an lvalue. ISO C prohibits
2435 qualified @code{void} return types on function definitions, so such
2436 return types always receive a warning even without this option.
2438 For C++, a function without return type always produces a diagnostic
2439 message, even when @option{-Wno-return-type} is specified. The only
2440 exceptions are @samp{main} and functions defined in system headers.
2444 Warn whenever a @code{switch} statement has an index of enumerated type
2445 and lacks a @code{case} for one or more of the named codes of that
2446 enumeration. (The presence of a @code{default} label prevents this
2447 warning.) @code{case} labels outside the enumeration range also
2448 provoke warnings when this option is used.
2450 @item -Wswitch-default
2451 @opindex Wswitch-switch
2452 Warn whenever a @code{switch} statement does not have a @code{default}
2456 @opindex Wswitch-enum
2457 Warn whenever a @code{switch} statement has an index of enumerated type
2458 and lacks a @code{case} for one or more of the named codes of that
2459 enumeration. @code{case} labels outside the enumeration range also
2460 provoke warnings when this option is used.
2464 Warn if any trigraphs are encountered that might change the meaning of
2465 the program (trigraphs within comments are not warned about).
2467 @item -Wunused-function
2468 @opindex Wunused-function
2469 Warn whenever a static function is declared but not defined or a
2470 non\-inline static function is unused.
2472 @item -Wunused-label
2473 @opindex Wunused-label
2474 Warn whenever a label is declared but not used.
2476 To suppress this warning use the @samp{unused} attribute
2477 (@pxref{Variable Attributes}).
2479 @item -Wunused-parameter
2480 @opindex Wunused-parameter
2481 Warn whenever a function parameter is unused aside from its declaration.
2483 To suppress this warning use the @samp{unused} attribute
2484 (@pxref{Variable Attributes}).
2486 @item -Wunused-variable
2487 @opindex Wunused-variable
2488 Warn whenever a local variable or non-constant static variable is unused
2489 aside from its declaration
2491 To suppress this warning use the @samp{unused} attribute
2492 (@pxref{Variable Attributes}).
2494 @item -Wunused-value
2495 @opindex Wunused-value
2496 Warn whenever a statement computes a result that is explicitly not used.
2498 To suppress this warning cast the expression to @samp{void}.
2502 All the above @option{-Wunused} options combined.
2504 In order to get a warning about an unused function parameter, you must
2505 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2506 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2508 @item -Wuninitialized
2509 @opindex Wuninitialized
2510 Warn if an automatic variable is used without first being initialized or
2511 if a variable may be clobbered by a @code{setjmp} call.
2513 These warnings are possible only in optimizing compilation,
2514 because they require data flow information that is computed only
2515 when optimizing. If you don't specify @option{-O}, you simply won't
2518 If you want to warn about code which uses the uninitialized value of the
2519 variable in its own initializer, use the @option{-Winit-self} option.
2521 These warnings occur only for variables that are candidates for
2522 register allocation. Therefore, they do not occur for a variable that
2523 is declared @code{volatile}, or whose address is taken, or whose size
2524 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2525 structures, unions or arrays, even when they are in registers.
2527 Note that there may be no warning about a variable that is used only
2528 to compute a value that itself is never used, because such
2529 computations may be deleted by data flow analysis before the warnings
2532 These warnings are made optional because GCC is not smart
2533 enough to see all the reasons why the code might be correct
2534 despite appearing to have an error. Here is one example of how
2555 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2556 always initialized, but GCC doesn't know this. Here is
2557 another common case:
2562 if (change_y) save_y = y, y = new_y;
2564 if (change_y) y = save_y;
2569 This has no bug because @code{save_y} is used only if it is set.
2571 @cindex @code{longjmp} warnings
2572 This option also warns when a non-volatile automatic variable might be
2573 changed by a call to @code{longjmp}. These warnings as well are possible
2574 only in optimizing compilation.
2576 The compiler sees only the calls to @code{setjmp}. It cannot know
2577 where @code{longjmp} will be called; in fact, a signal handler could
2578 call it at any point in the code. As a result, you may get a warning
2579 even when there is in fact no problem because @code{longjmp} cannot
2580 in fact be called at the place which would cause a problem.
2582 Some spurious warnings can be avoided if you declare all the functions
2583 you use that never return as @code{noreturn}. @xref{Function
2586 @item -Wunknown-pragmas
2587 @opindex Wunknown-pragmas
2588 @cindex warning for unknown pragmas
2589 @cindex unknown pragmas, warning
2590 @cindex pragmas, warning of unknown
2591 Warn when a #pragma directive is encountered which is not understood by
2592 GCC@. If this command line option is used, warnings will even be issued
2593 for unknown pragmas in system header files. This is not the case if
2594 the warnings were only enabled by the @option{-Wall} command line option.
2596 @item -Wstrict-aliasing
2597 @opindex Wstrict-aliasing
2598 This option is only active when @option{-fstrict-aliasing} is active.
2599 It warns about code which might break the strict aliasing rules that the
2600 compiler is using for optimization. The warning does not catch all
2601 cases, but does attempt to catch the more common pitfalls. It is
2602 included in @option{-Wall}.
2604 @item -Wstrict-aliasing=2
2605 @opindex Wstrict-aliasing=2
2606 This option is only active when @option{-fstrict-aliasing} is active.
2607 It warns about all code which might break the strict aliasing rules that the
2608 compiler is using for optimization. This warning catches all cases, but
2609 it will also give a warning for some ambiguous cases that are safe.
2613 All of the above @samp{-W} options combined. This enables all the
2614 warnings about constructions that some users consider questionable, and
2615 that are easy to avoid (or modify to prevent the warning), even in
2616 conjunction with macros. This also enables some language-specific
2617 warnings described in @ref{C++ Dialect Options} and
2618 @ref{Objective-C and Objective-C++ Dialect Options}.
2621 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2622 Some of them warn about constructions that users generally do not
2623 consider questionable, but which occasionally you might wish to check
2624 for; others warn about constructions that are necessary or hard to avoid
2625 in some cases, and there is no simple way to modify the code to suppress
2632 (This option used to be called @option{-W}. The older name is still
2633 supported, but the newer name is more descriptive.) Print extra warning
2634 messages for these events:
2638 A function can return either with or without a value. (Falling
2639 off the end of the function body is considered returning without
2640 a value.) For example, this function would evoke such a
2654 An expression-statement or the left-hand side of a comma expression
2655 contains no side effects.
2656 To suppress the warning, cast the unused expression to void.
2657 For example, an expression such as @samp{x[i,j]} will cause a warning,
2658 but @samp{x[(void)i,j]} will not.
2661 An unsigned value is compared against zero with @samp{<} or @samp{>=}.
2664 Storage-class specifiers like @code{static} are not the first things in
2665 a declaration. According to the C Standard, this usage is obsolescent.
2668 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2672 A comparison between signed and unsigned values could produce an
2673 incorrect result when the signed value is converted to unsigned.
2674 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2677 An aggregate has an initializer which does not initialize all members.
2678 This warning can be independently controlled by
2679 @option{-Wmissing-field-initializers}.
2682 A function parameter is declared without a type specifier in K&R-style
2690 An empty body occurs in an @samp{if} or @samp{else} statement.
2693 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2694 @samp{>}, or @samp{>=}.
2697 A variable might be changed by @samp{longjmp} or @samp{vfork}.
2700 Any of several floating-point events that often indicate errors, such as
2701 overflow, underflow, loss of precision, etc.
2703 @item @r{(C++ only)}
2704 An enumerator and a non-enumerator both appear in a conditional expression.
2706 @item @r{(C++ only)}
2707 A non-static reference or non-static @samp{const} member appears in a
2708 class without constructors.
2710 @item @r{(C++ only)}
2711 Ambiguous virtual bases.
2713 @item @r{(C++ only)}
2714 Subscripting an array which has been declared @samp{register}.
2716 @item @r{(C++ only)}
2717 Taking the address of a variable which has been declared @samp{register}.
2719 @item @r{(C++ only)}
2720 A base class is not initialized in a derived class' copy constructor.
2723 @item -Wno-div-by-zero
2724 @opindex Wno-div-by-zero
2725 @opindex Wdiv-by-zero
2726 Do not warn about compile-time integer division by zero. Floating point
2727 division by zero is not warned about, as it can be a legitimate way of
2728 obtaining infinities and NaNs.
2730 @item -Wsystem-headers
2731 @opindex Wsystem-headers
2732 @cindex warnings from system headers
2733 @cindex system headers, warnings from
2734 Print warning messages for constructs found in system header files.
2735 Warnings from system headers are normally suppressed, on the assumption
2736 that they usually do not indicate real problems and would only make the
2737 compiler output harder to read. Using this command line option tells
2738 GCC to emit warnings from system headers as if they occurred in user
2739 code. However, note that using @option{-Wall} in conjunction with this
2740 option will @emph{not} warn about unknown pragmas in system
2741 headers---for that, @option{-Wunknown-pragmas} must also be used.
2744 @opindex Wfloat-equal
2745 Warn if floating point values are used in equality comparisons.
2747 The idea behind this is that sometimes it is convenient (for the
2748 programmer) to consider floating-point values as approximations to
2749 infinitely precise real numbers. If you are doing this, then you need
2750 to compute (by analyzing the code, or in some other way) the maximum or
2751 likely maximum error that the computation introduces, and allow for it
2752 when performing comparisons (and when producing output, but that's a
2753 different problem). In particular, instead of testing for equality, you
2754 would check to see whether the two values have ranges that overlap; and
2755 this is done with the relational operators, so equality comparisons are
2758 @item -Wtraditional @r{(C only)}
2759 @opindex Wtraditional
2760 Warn about certain constructs that behave differently in traditional and
2761 ISO C@. Also warn about ISO C constructs that have no traditional C
2762 equivalent, and/or problematic constructs which should be avoided.
2766 Macro parameters that appear within string literals in the macro body.
2767 In traditional C macro replacement takes place within string literals,
2768 but does not in ISO C@.
2771 In traditional C, some preprocessor directives did not exist.
2772 Traditional preprocessors would only consider a line to be a directive
2773 if the @samp{#} appeared in column 1 on the line. Therefore
2774 @option{-Wtraditional} warns about directives that traditional C
2775 understands but would ignore because the @samp{#} does not appear as the
2776 first character on the line. It also suggests you hide directives like
2777 @samp{#pragma} not understood by traditional C by indenting them. Some
2778 traditional implementations would not recognize @samp{#elif}, so it
2779 suggests avoiding it altogether.
2782 A function-like macro that appears without arguments.
2785 The unary plus operator.
2788 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2789 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2790 constants.) Note, these suffixes appear in macros defined in the system
2791 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2792 Use of these macros in user code might normally lead to spurious
2793 warnings, however GCC's integrated preprocessor has enough context to
2794 avoid warning in these cases.
2797 A function declared external in one block and then used after the end of
2801 A @code{switch} statement has an operand of type @code{long}.
2804 A non-@code{static} function declaration follows a @code{static} one.
2805 This construct is not accepted by some traditional C compilers.
2808 The ISO type of an integer constant has a different width or
2809 signedness from its traditional type. This warning is only issued if
2810 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2811 typically represent bit patterns, are not warned about.
2814 Usage of ISO string concatenation is detected.
2817 Initialization of automatic aggregates.
2820 Identifier conflicts with labels. Traditional C lacks a separate
2821 namespace for labels.
2824 Initialization of unions. If the initializer is zero, the warning is
2825 omitted. This is done under the assumption that the zero initializer in
2826 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2827 initializer warnings and relies on default initialization to zero in the
2831 Conversions by prototypes between fixed/floating point values and vice
2832 versa. The absence of these prototypes when compiling with traditional
2833 C would cause serious problems. This is a subset of the possible
2834 conversion warnings, for the full set use @option{-Wconversion}.
2837 Use of ISO C style function definitions. This warning intentionally is
2838 @emph{not} issued for prototype declarations or variadic functions
2839 because these ISO C features will appear in your code when using
2840 libiberty's traditional C compatibility macros, @code{PARAMS} and
2841 @code{VPARAMS}. This warning is also bypassed for nested functions
2842 because that feature is already a GCC extension and thus not relevant to
2843 traditional C compatibility.
2846 @item -Wdeclaration-after-statement @r{(C only)}
2847 @opindex Wdeclaration-after-statement
2848 Warn when a declaration is found after a statement in a block. This
2849 construct, known from C++, was introduced with ISO C99 and is by default
2850 allowed in GCC@. It is not supported by ISO C90 and was not supported by
2851 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
2855 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2857 @item -Wendif-labels
2858 @opindex Wendif-labels
2859 Warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
2863 Warn whenever a local variable shadows another local variable, parameter or
2864 global variable or whenever a built-in function is shadowed.
2866 @item -Wlarger-than-@var{len}
2867 @opindex Wlarger-than
2868 Warn whenever an object of larger than @var{len} bytes is defined.
2870 @item -Wpointer-arith
2871 @opindex Wpointer-arith
2872 Warn about anything that depends on the ``size of'' a function type or
2873 of @code{void}. GNU C assigns these types a size of 1, for
2874 convenience in calculations with @code{void *} pointers and pointers
2877 @item -Wbad-function-cast @r{(C only)}
2878 @opindex Wbad-function-cast
2879 Warn whenever a function call is cast to a non-matching type.
2880 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2884 Warn whenever a pointer is cast so as to remove a type qualifier from
2885 the target type. For example, warn if a @code{const char *} is cast
2886 to an ordinary @code{char *}.
2889 @opindex Wcast-align
2890 Warn whenever a pointer is cast such that the required alignment of the
2891 target is increased. For example, warn if a @code{char *} is cast to
2892 an @code{int *} on machines where integers can only be accessed at
2893 two- or four-byte boundaries.
2895 @item -Wwrite-strings
2896 @opindex Wwrite-strings
2897 When compiling C, give string constants the type @code{const
2898 char[@var{length}]} so that
2899 copying the address of one into a non-@code{const} @code{char *}
2900 pointer will get a warning; when compiling C++, warn about the
2901 deprecated conversion from string constants to @code{char *}.
2902 These warnings will help you find at
2903 compile time code that can try to write into a string constant, but
2904 only if you have been very careful about using @code{const} in
2905 declarations and prototypes. Otherwise, it will just be a nuisance;
2906 this is why we did not make @option{-Wall} request these warnings.
2909 @opindex Wconversion
2910 Warn if a prototype causes a type conversion that is different from what
2911 would happen to the same argument in the absence of a prototype. This
2912 includes conversions of fixed point to floating and vice versa, and
2913 conversions changing the width or signedness of a fixed point argument
2914 except when the same as the default promotion.
2916 Also, warn if a negative integer constant expression is implicitly
2917 converted to an unsigned type. For example, warn about the assignment
2918 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2919 casts like @code{(unsigned) -1}.
2921 @item -Wsign-compare
2922 @opindex Wsign-compare
2923 @cindex warning for comparison of signed and unsigned values
2924 @cindex comparison of signed and unsigned values, warning
2925 @cindex signed and unsigned values, comparison warning
2926 Warn when a comparison between signed and unsigned values could produce
2927 an incorrect result when the signed value is converted to unsigned.
2928 This warning is also enabled by @option{-Wextra}; to get the other warnings
2929 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
2931 @item -Waggregate-return
2932 @opindex Waggregate-return
2933 Warn if any functions that return structures or unions are defined or
2934 called. (In languages where you can return an array, this also elicits
2937 @item -Wstrict-prototypes @r{(C only)}
2938 @opindex Wstrict-prototypes
2939 Warn if a function is declared or defined without specifying the
2940 argument types. (An old-style function definition is permitted without
2941 a warning if preceded by a declaration which specifies the argument
2944 @item -Wold-style-definition @r{(C only)}
2945 @opindex Wold-style-definition
2946 Warn if an old-style function definition is used. A warning is given
2947 even if there is a previous prototype.
2949 @item -Wmissing-prototypes @r{(C only)}
2950 @opindex Wmissing-prototypes
2951 Warn if a global function is defined without a previous prototype
2952 declaration. This warning is issued even if the definition itself
2953 provides a prototype. The aim is to detect global functions that fail
2954 to be declared in header files.
2956 @item -Wmissing-declarations @r{(C only)}
2957 @opindex Wmissing-declarations
2958 Warn if a global function is defined without a previous declaration.
2959 Do so even if the definition itself provides a prototype.
2960 Use this option to detect global functions that are not declared in
2963 @item -Wmissing-field-initializers
2964 @opindex Wmissing-field-initializers
2967 Warn if a structure's initializer has some fields missing. For
2968 example, the following code would cause such a warning, because
2969 @code{x.h} is implicitly zero:
2972 struct s @{ int f, g, h; @};
2973 struct s x = @{ 3, 4 @};
2976 This option does not warn about designated initializers, so the following
2977 modification would not trigger a warning:
2980 struct s @{ int f, g, h; @};
2981 struct s x = @{ .f = 3, .g = 4 @};
2984 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
2985 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
2987 @item -Wmissing-noreturn
2988 @opindex Wmissing-noreturn
2989 Warn about functions which might be candidates for attribute @code{noreturn}.
2990 Note these are only possible candidates, not absolute ones. Care should
2991 be taken to manually verify functions actually do not ever return before
2992 adding the @code{noreturn} attribute, otherwise subtle code generation
2993 bugs could be introduced. You will not get a warning for @code{main} in
2994 hosted C environments.
2996 @item -Wmissing-format-attribute
2997 @opindex Wmissing-format-attribute
2999 If @option{-Wformat} is enabled, also warn about functions which might be
3000 candidates for @code{format} attributes. Note these are only possible
3001 candidates, not absolute ones. GCC will guess that @code{format}
3002 attributes might be appropriate for any function that calls a function
3003 like @code{vprintf} or @code{vscanf}, but this might not always be the
3004 case, and some functions for which @code{format} attributes are
3005 appropriate may not be detected. This option has no effect unless
3006 @option{-Wformat} is enabled (possibly by @option{-Wall}).
3008 @item -Wno-multichar
3009 @opindex Wno-multichar
3011 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
3012 Usually they indicate a typo in the user's code, as they have
3013 implementation-defined values, and should not be used in portable code.
3015 @item -Wno-deprecated-declarations
3016 @opindex Wno-deprecated-declarations
3017 Do not warn about uses of functions, variables, and types marked as
3018 deprecated by using the @code{deprecated} attribute.
3019 (@pxref{Function Attributes}, @pxref{Variable Attributes},
3020 @pxref{Type Attributes}.)
3024 Warn if a structure is given the packed attribute, but the packed
3025 attribute has no effect on the layout or size of the structure.
3026 Such structures may be mis-aligned for little benefit. For
3027 instance, in this code, the variable @code{f.x} in @code{struct bar}
3028 will be misaligned even though @code{struct bar} does not itself
3029 have the packed attribute:
3036 @} __attribute__((packed));
3046 Warn if padding is included in a structure, either to align an element
3047 of the structure or to align the whole structure. Sometimes when this
3048 happens it is possible to rearrange the fields of the structure to
3049 reduce the padding and so make the structure smaller.
3051 @item -Wredundant-decls
3052 @opindex Wredundant-decls
3053 Warn if anything is declared more than once in the same scope, even in
3054 cases where multiple declaration is valid and changes nothing.
3056 @item -Wnested-externs @r{(C only)}
3057 @opindex Wnested-externs
3058 Warn if an @code{extern} declaration is encountered within a function.
3060 @item -Wunreachable-code
3061 @opindex Wunreachable-code
3062 Warn if the compiler detects that code will never be executed.
3064 This option is intended to warn when the compiler detects that at
3065 least a whole line of source code will never be executed, because
3066 some condition is never satisfied or because it is after a
3067 procedure that never returns.
3069 It is possible for this option to produce a warning even though there
3070 are circumstances under which part of the affected line can be executed,
3071 so care should be taken when removing apparently-unreachable code.
3073 For instance, when a function is inlined, a warning may mean that the
3074 line is unreachable in only one inlined copy of the function.
3076 This option is not made part of @option{-Wall} because in a debugging
3077 version of a program there is often substantial code which checks
3078 correct functioning of the program and is, hopefully, unreachable
3079 because the program does work. Another common use of unreachable
3080 code is to provide behavior which is selectable at compile-time.
3084 Warn if a function can not be inlined and it was declared as inline.
3085 Even with this option, the compiler will not warn about failures to
3086 inline functions declared in system headers.
3088 The compiler uses a variety of heuristics to determine whether or not
3089 to inline a function. For example, the compiler takes into account
3090 the size of the function being inlined and the the amount of inlining
3091 that has already been done in the current function. Therefore,
3092 seemingly insignificant changes in the source program can cause the
3093 warnings produced by @option{-Winline} to appear or disappear.
3095 @item -Wno-invalid-offsetof @r{(C++ only)}
3096 @opindex Wno-invalid-offsetof
3097 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
3098 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
3099 to a non-POD type is undefined. In existing C++ implementations,
3100 however, @samp{offsetof} typically gives meaningful results even when
3101 applied to certain kinds of non-POD types. (Such as a simple
3102 @samp{struct} that fails to be a POD type only by virtue of having a
3103 constructor.) This flag is for users who are aware that they are
3104 writing nonportable code and who have deliberately chosen to ignore the
3107 The restrictions on @samp{offsetof} may be relaxed in a future version
3108 of the C++ standard.
3111 @opindex Winvalid-pch
3112 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
3113 the search path but can't be used.
3117 @opindex Wno-long-long
3118 Warn if @samp{long long} type is used. This is default. To inhibit
3119 the warning messages, use @option{-Wno-long-long}. Flags
3120 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
3121 only when @option{-pedantic} flag is used.
3123 @item -Wvariadic-macros
3124 @opindex Wvariadic-macros
3125 @opindex Wno-variadic-macros
3126 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
3127 alternate syntax when in pedantic ISO C99 mode. This is default.
3128 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
3130 @item -Wdisabled-optimization
3131 @opindex Wdisabled-optimization
3132 Warn if a requested optimization pass is disabled. This warning does
3133 not generally indicate that there is anything wrong with your code; it
3134 merely indicates that GCC's optimizers were unable to handle the code
3135 effectively. Often, the problem is that your code is too big or too
3136 complex; GCC will refuse to optimize programs when the optimization
3137 itself is likely to take inordinate amounts of time.
3139 @item -Wno-pointer-sign
3140 @opindex Wno-pointer-sign
3141 Don't warn for pointer argument passing or assignment with different signedness.
3142 Only useful in the negative form since this warning is enabled by default.
3143 This option is only supported for C and Objective-C@.
3147 Make all warnings into errors.
3150 @node Debugging Options
3151 @section Options for Debugging Your Program or GCC
3152 @cindex options, debugging
3153 @cindex debugging information options
3155 GCC has various special options that are used for debugging
3156 either your program or GCC:
3161 Produce debugging information in the operating system's native format
3162 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
3165 On most systems that use stabs format, @option{-g} enables use of extra
3166 debugging information that only GDB can use; this extra information
3167 makes debugging work better in GDB but will probably make other debuggers
3169 refuse to read the program. If you want to control for certain whether
3170 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
3171 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
3173 GCC allows you to use @option{-g} with
3174 @option{-O}. The shortcuts taken by optimized code may occasionally
3175 produce surprising results: some variables you declared may not exist
3176 at all; flow of control may briefly move where you did not expect it;
3177 some statements may not be executed because they compute constant
3178 results or their values were already at hand; some statements may
3179 execute in different places because they were moved out of loops.
3181 Nevertheless it proves possible to debug optimized output. This makes
3182 it reasonable to use the optimizer for programs that might have bugs.
3184 The following options are useful when GCC is generated with the
3185 capability for more than one debugging format.
3189 Produce debugging information for use by GDB@. This means to use the
3190 most expressive format available (DWARF 2, stabs, or the native format
3191 if neither of those are supported), including GDB extensions if at all
3196 Produce debugging information in stabs format (if that is supported),
3197 without GDB extensions. This is the format used by DBX on most BSD
3198 systems. On MIPS, Alpha and System V Release 4 systems this option
3199 produces stabs debugging output which is not understood by DBX or SDB@.
3200 On System V Release 4 systems this option requires the GNU assembler.
3202 @item -feliminate-unused-debug-symbols
3203 @opindex feliminate-unused-debug-symbols
3204 Produce debugging information in stabs format (if that is supported),
3205 for only symbols that are actually used.
3209 Produce debugging information in stabs format (if that is supported),
3210 using GNU extensions understood only by the GNU debugger (GDB)@. The
3211 use of these extensions is likely to make other debuggers crash or
3212 refuse to read the program.
3216 Produce debugging information in COFF format (if that is supported).
3217 This is the format used by SDB on most System V systems prior to
3222 Produce debugging information in XCOFF format (if that is supported).
3223 This is the format used by the DBX debugger on IBM RS/6000 systems.
3227 Produce debugging information in XCOFF format (if that is supported),
3228 using GNU extensions understood only by the GNU debugger (GDB)@. The
3229 use of these extensions is likely to make other debuggers crash or
3230 refuse to read the program, and may cause assemblers other than the GNU
3231 assembler (GAS) to fail with an error.
3235 Produce debugging information in DWARF version 2 format (if that is
3236 supported). This is the format used by DBX on IRIX 6. With this
3237 option, GCC uses features of DWARF version 3 when they are useful;
3238 version 3 is upward compatible with version 2, but may still cause
3239 problems for older debuggers.
3243 Produce debugging information in VMS debug format (if that is
3244 supported). This is the format used by DEBUG on VMS systems.
3247 @itemx -ggdb@var{level}
3248 @itemx -gstabs@var{level}
3249 @itemx -gcoff@var{level}
3250 @itemx -gxcoff@var{level}
3251 @itemx -gvms@var{level}
3252 Request debugging information and also use @var{level} to specify how
3253 much information. The default level is 2.
3255 Level 1 produces minimal information, enough for making backtraces in
3256 parts of the program that you don't plan to debug. This includes
3257 descriptions of functions and external variables, but no information
3258 about local variables and no line numbers.
3260 Level 3 includes extra information, such as all the macro definitions
3261 present in the program. Some debuggers support macro expansion when
3262 you use @option{-g3}.
3264 @option{-gdwarf-2} does not accept a concatenated debug level, because
3265 GCC used to support an option @option{-gdwarf} that meant to generate
3266 debug information in version 1 of the DWARF format (which is very
3267 different from version 2), and it would have been too confusing. That
3268 debug format is long obsolete, but the option cannot be changed now.
3269 Instead use an additional @option{-g@var{level}} option to change the
3270 debug level for DWARF2.
3272 @item -feliminate-dwarf2-dups
3273 @opindex feliminate-dwarf2-dups
3274 Compress DWARF2 debugging information by eliminating duplicated
3275 information about each symbol. This option only makes sense when
3276 generating DWARF2 debugging information with @option{-gdwarf-2}.
3278 @cindex @command{prof}
3281 Generate extra code to write profile information suitable for the
3282 analysis program @command{prof}. You must use this option when compiling
3283 the source files you want data about, and you must also use it when
3286 @cindex @command{gprof}
3289 Generate extra code to write profile information suitable for the
3290 analysis program @command{gprof}. You must use this option when compiling
3291 the source files you want data about, and you must also use it when
3296 Makes the compiler print out each function name as it is compiled, and
3297 print some statistics about each pass when it finishes.
3300 @opindex ftime-report
3301 Makes the compiler print some statistics about the time consumed by each
3302 pass when it finishes.
3305 @opindex fmem-report
3306 Makes the compiler print some statistics about permanent memory
3307 allocation when it finishes.
3309 @item -fprofile-arcs
3310 @opindex fprofile-arcs
3311 Add code so that program flow @dfn{arcs} are instrumented. During
3312 execution the program records how many times each branch and call is
3313 executed and how many times it is taken or returns. When the compiled
3314 program exits it saves this data to a file called
3315 @file{@var{auxname}.gcda} for each source file. The data may be used for
3316 profile-directed optimizations (@option{-fbranch-probabilities}), or for
3317 test coverage analysis (@option{-ftest-coverage}). Each object file's
3318 @var{auxname} is generated from the name of the output file, if
3319 explicitly specified and it is not the final executable, otherwise it is
3320 the basename of the source file. In both cases any suffix is removed
3321 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
3322 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
3327 Compile the source files with @option{-fprofile-arcs} plus optimization
3328 and code generation options. For test coverage analysis, use the
3329 additional @option{-ftest-coverage} option. You do not need to profile
3330 every source file in a program.
3333 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
3334 (the latter implies the former).
3337 Run the program on a representative workload to generate the arc profile
3338 information. This may be repeated any number of times. You can run
3339 concurrent instances of your program, and provided that the file system
3340 supports locking, the data files will be correctly updated. Also
3341 @code{fork} calls are detected and correctly handled (double counting
3345 For profile-directed optimizations, compile the source files again with
3346 the same optimization and code generation options plus
3347 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
3348 Control Optimization}).
3351 For test coverage analysis, use @command{gcov} to produce human readable
3352 information from the @file{.gcno} and @file{.gcda} files. Refer to the
3353 @command{gcov} documentation for further information.
3357 With @option{-fprofile-arcs}, for each function of your program GCC
3358 creates a program flow graph, then finds a spanning tree for the graph.
3359 Only arcs that are not on the spanning tree have to be instrumented: the
3360 compiler adds code to count the number of times that these arcs are
3361 executed. When an arc is the only exit or only entrance to a block, the
3362 instrumentation code can be added to the block; otherwise, a new basic
3363 block must be created to hold the instrumentation code.
3365 @item -ftree-based-profiling
3366 @opindex ftree-based-profiling
3367 This option is used in addition to @option{-fprofile-arcs} or
3368 @option{-fbranch-probabilities} to control whether those optimizations
3369 are performed on a tree-based or rtl-based internal representation.
3370 If you use this option when compiling with @option{-fprofile-arcs},
3371 you must also use it when compiling later with @option{-fbranch-probabilities}.
3372 Currently the tree-based optimization is in an early stage of
3373 development, and this option is recommended only for those people
3374 working on improving it.
3377 @item -ftest-coverage
3378 @opindex ftest-coverage
3379 Produce a notes file that the @command{gcov} code-coverage utility
3380 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
3381 show program coverage. Each source file's note file is called
3382 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
3383 above for a description of @var{auxname} and instructions on how to
3384 generate test coverage data. Coverage data will match the source files
3385 more closely, if you do not optimize.
3387 @item -d@var{letters}
3388 @item -fdump-rtl-@var{pass}
3390 Says to make debugging dumps during compilation at times specified by
3391 @var{letters}. This is used for debugging the RTL-based passes of the
3392 compiler. The file names for most of the dumps are made by appending a
3393 pass number and a word to the @var{dumpname}. @var{dumpname} is generated
3394 from the name of the output file, if explicitly specified and it is not
3395 an executable, otherwise it is the basename of the source file.
3397 Most debug dumps can be enabled either passing a letter to the @option{-d}
3398 option, or with a long @option{-fdump-rtl} switch; here are the possible
3399 letters for use in @var{letters} and @var{pass}, and their meanings:
3404 Annotate the assembler output with miscellaneous debugging information.
3407 @itemx -fdump-rtl-bp
3409 @opindex fdump-rtl-bp
3410 Dump after computing branch probabilities, to @file{@var{file}.09.bp}.
3413 @itemx -fdump-rtl-bbro
3415 @opindex fdump-rtl-bbro
3416 Dump after block reordering, to @file{@var{file}.30.bbro}.
3419 @itemx -fdump-rtl-combine
3421 @opindex fdump-rtl-combine
3422 Dump after instruction combination, to the file @file{@var{file}.17.combine}.
3425 @itemx -fdump-rtl-ce1
3426 @itemx -fdump-rtl-ce2
3428 @opindex fdump-rtl-ce1
3429 @opindex fdump-rtl-ce2
3430 @option{-dC} and @option{-fdump-rtl-ce1} enable dumping after the
3431 first if conversion, to the file @file{@var{file}.11.ce1}. @option{-dC}
3432 and @option{-fdump-rtl-ce2} enable dumping after the second if
3433 conversion, to the file @file{@var{file}.18.ce2}.
3436 @itemx -fdump-rtl-btl
3437 @itemx -fdump-rtl-dbr
3439 @opindex fdump-rtl-btl
3440 @opindex fdump-rtl-dbr
3441 @option{-dd} and @option{-fdump-rtl-btl} enable dumping after branch
3442 target load optimization, to to @file{@var{file}.31.btl}. @option{-dd}
3443 and @option{-fdump-rtl-dbr} enable dumping after delayed branch
3444 scheduling, to @file{@var{file}.36.dbr}.
3448 Dump all macro definitions, at the end of preprocessing, in addition to
3452 @itemx -fdump-rtl-ce3
3454 @opindex fdump-rtl-ce3
3455 Dump after the third if conversion, to @file{@var{file}.28.ce3}.
3458 @itemx -fdump-rtl-cfg
3459 @itemx -fdump-rtl-life
3461 @opindex fdump-rtl-cfg
3462 @opindex fdump-rtl-life
3463 @option{-df} and @option{-fdump-rtl-cfg} enable dumping after control
3464 and data flow analysis, to @file{@var{file}.08.cfg}. @option{-df}
3465 and @option{-fdump-rtl-cfg} enable dumping dump after life analysis,
3466 to @file{@var{file}.16.life}.
3469 @itemx -fdump-rtl-greg
3471 @opindex fdump-rtl-greg
3472 Dump after global register allocation, to @file{@var{file}.23.greg}.
3475 @itemx -fdump-rtl-gcse
3476 @itemx -fdump-rtl-bypass
3478 @opindex fdump-rtl-gcse
3479 @opindex fdump-rtl-bypass
3480 @option{-dG} and @option{-fdump-rtl-gcse} enable dumping after GCSE, to
3481 @file{@var{file}.05.gcse}. @option{-dG} and @option{-fdump-rtl-bypass}
3482 enable dumping after jump bypassing and control flow optimizations, to
3483 @file{@var{file}.07.bypass}.
3486 @itemx -fdump-rtl-eh
3488 @opindex fdump-rtl-eh
3489 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
3492 @itemx -fdump-rtl-sibling
3494 @opindex fdump-rtl-sibling
3495 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
3498 @itemx -fdump-rtl-jump
3500 @opindex fdump-rtl-jump
3501 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
3504 @itemx -fdump-rtl-stack
3506 @opindex fdump-rtl-stack
3507 Dump after conversion from registers to stack, to @file{@var{file}.33.stack}.
3510 @itemx -fdump-rtl-lreg
3512 @opindex fdump-rtl-lreg
3513 Dump after local register allocation, to @file{@var{file}.22.lreg}.
3516 @itemx -fdump-rtl-loop
3517 @itemx -fdump-rtl-loop2
3519 @opindex fdump-rtl-loop
3520 @opindex fdump-rtl-loop2
3521 @option{-dL} and @option{-fdump-rtl-loop} enable dumping after the first
3522 loop optimization pass, to @file{@var{file}.06.loop}. @option{-dL} and
3523 @option{-fdump-rtl-loop2} enable dumping after the second pass, to
3524 @file{@var{file}.13.loop2}.
3527 @itemx -fdump-rtl-sms
3529 @opindex fdump-rtl-sms
3530 Dump after modulo scheduling, to @file{@var{file}.20.sms}.
3533 @itemx -fdump-rtl-mach
3535 @opindex fdump-rtl-mach
3536 Dump after performing the machine dependent reorganization pass, to
3537 @file{@var{file}.35.mach}.
3540 @itemx -fdump-rtl-rnreg
3542 @opindex fdump-rtl-rnreg
3543 Dump after register renumbering, to @file{@var{file}.29.rnreg}.
3546 @itemx -fdump-rtl-regmove
3548 @opindex fdump-rtl-regmove
3549 Dump after the register move pass, to @file{@var{file}.19.regmove}.
3552 @itemx -fdump-rtl-postreload
3554 @opindex fdump-rtl-postreload
3555 Dump after post-reload optimizations, to @file{@var{file}.24.postreload}.
3558 @itemx -fdump-rtl-expand
3560 @opindex fdump-rtl-expand
3561 Dump after RTL generation, to @file{@var{file}.00.expand}.
3564 @itemx -fdump-rtl-sched2
3566 @opindex fdump-rtl-sched2
3567 Dump after the second scheduling pass, to @file{@var{file}.32.sched2}.
3570 @itemx -fdump-rtl-cse
3572 @opindex fdump-rtl-cse
3573 Dump after CSE (including the jump optimization that sometimes follows
3574 CSE), to @file{@var{file}.04.cse}.
3577 @itemx -fdump-rtl-sched
3579 @opindex fdump-rtl-sched
3580 Dump after the first scheduling pass, to @file{@var{file}.21.sched}.
3583 @itemx -fdump-rtl-cse2
3585 @opindex fdump-rtl-cse2
3586 Dump after the second CSE pass (including the jump optimization that
3587 sometimes follows CSE), to @file{@var{file}.15.cse2}.
3590 @itemx -fdump-rtl-tracer
3592 @opindex fdump-rtl-tracer
3593 Dump after running tracer, to @file{@var{file}.12.tracer}.
3596 @itemx -fdump-rtl-vpt
3597 @itemx -fdump-rtl-vartrack
3599 @opindex fdump-rtl-vpt
3600 @opindex fdump-rtl-vartrack
3601 @option{-dV} and @option{-fdump-rtl-vpt} enable dumping after the value
3602 profile transformations, to @file{@var{file}.10.vpt}. @option{-dV}
3603 and @option{-fdump-rtl-vartrack} enable dumping after variable tracking,
3604 to @file{@var{file}.34.vartrack}.
3607 @itemx -fdump-rtl-flow2
3609 @opindex fdump-rtl-flow2
3610 Dump after the second flow pass, to @file{@var{file}.26.flow2}.
3613 @itemx -fdump-rtl-peephole2
3615 @opindex fdump-rtl-peephole2
3616 Dump after the peephole pass, to @file{@var{file}.27.peephole2}.
3619 @itemx -fdump-rtl-web
3621 @opindex fdump-rtl-web
3622 Dump after live range splitting, to @file{@var{file}.14.web}.
3625 @itemx -fdump-rtl-all
3627 @opindex fdump-rtl-all
3628 Produce all the dumps listed above.
3632 Produce a core dump whenever an error occurs.
3636 Print statistics on memory usage, at the end of the run, to
3641 Annotate the assembler output with a comment indicating which
3642 pattern and alternative was used. The length of each instruction is
3647 Dump the RTL in the assembler output as a comment before each instruction.
3648 Also turns on @option{-dp} annotation.
3652 For each of the other indicated dump files (either with @option{-d} or
3653 @option{-fdump-rtl-@var{pass}}), dump a representation of the control flow
3654 graph suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3658 Just generate RTL for a function instead of compiling it. Usually used
3659 with @samp{r} (@option{-fdump-rtl-expand}).
3663 Dump debugging information during parsing, to standard error.
3666 @item -fdump-unnumbered
3667 @opindex fdump-unnumbered
3668 When doing debugging dumps (see @option{-d} option above), suppress instruction
3669 numbers and line number note output. This makes it more feasible to
3670 use diff on debugging dumps for compiler invocations with different
3671 options, in particular with and without @option{-g}.
3673 @item -fdump-translation-unit @r{(C and C++ only)}
3674 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
3675 @opindex fdump-translation-unit
3676 Dump a representation of the tree structure for the entire translation
3677 unit to a file. The file name is made by appending @file{.tu} to the
3678 source file name. If the @samp{-@var{options}} form is used, @var{options}
3679 controls the details of the dump as described for the
3680 @option{-fdump-tree} options.
3682 @item -fdump-class-hierarchy @r{(C++ only)}
3683 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3684 @opindex fdump-class-hierarchy
3685 Dump a representation of each class's hierarchy and virtual function
3686 table layout to a file. The file name is made by appending @file{.class}
3687 to the source file name. If the @samp{-@var{options}} form is used,
3688 @var{options} controls the details of the dump as described for the
3689 @option{-fdump-tree} options.
3691 @item -fdump-ipa-@var{switch}
3693 Control the dumping at various stages of inter-procedural analysis
3694 language tree to a file. The file name is generated by appending a switch
3695 specific suffix to the source file name. The following dumps are possible:
3699 Enables all inter-procedural analysis dumps; currently the only produced
3700 dump is the @samp{cgraph} dump.
3703 Dumps information about call-graph optimization, unused function removal,
3704 and inlining decisions.
3707 @item -fdump-tree-@var{switch} @r{(C and C++ only)}
3708 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C and C++ only)}
3710 Control the dumping at various stages of processing the intermediate
3711 language tree to a file. The file name is generated by appending a switch
3712 specific suffix to the source file name. If the @samp{-@var{options}}
3713 form is used, @var{options} is a list of @samp{-} separated options that
3714 control the details of the dump. Not all options are applicable to all
3715 dumps, those which are not meaningful will be ignored. The following
3716 options are available
3720 Print the address of each node. Usually this is not meaningful as it
3721 changes according to the environment and source file. Its primary use
3722 is for tying up a dump file with a debug environment.
3724 Inhibit dumping of members of a scope or body of a function merely
3725 because that scope has been reached. Only dump such items when they
3726 are directly reachable by some other path. When dumping pretty-printed
3727 trees, this option inhibits dumping the bodies of control structures.
3729 Print a raw representation of the tree. By default, trees are
3730 pretty-printed into a C-like representation.
3732 Enable more detailed dumps (not honored by every dump option).
3734 Enable dumping various statistics about the pass (not honored by every dump
3737 Enable showing basic block boundaries (disabled in raw dumps).
3739 Enable showing virtual operands for every statement.
3741 Enable showing line numbers for statements.
3743 Enable showing the unique ID (@code{DECL_UID}) for each variable.
3745 Turn on all options, except @option{raw}, @option{slim} and @option{lineno}.
3748 The following tree dumps are possible:
3752 Dump before any tree based optimization, to @file{@var{file}.original}.
3755 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3758 Dump after function inlining, to @file{@var{file}.inlined}.
3761 @opindex fdump-tree-gimple
3762 Dump each function before and after the gimplification pass to a file. The
3763 file name is made by appending @file{.gimple} to the source file name.
3766 @opindex fdump-tree-cfg
3767 Dump the control flow graph of each function to a file. The file name is
3768 made by appending @file{.cfg} to the source file name.
3771 @opindex fdump-tree-vcg
3772 Dump the control flow graph of each function to a file in VCG format. The
3773 file name is made by appending @file{.vcg} to the source file name. Note
3774 that if the file contains more than one function, the generated file cannot
3775 be used directly by VCG@. You will need to cut and paste each function's
3776 graph into its own separate file first.
3779 @opindex fdump-tree-ch
3780 Dump each function after copying loop headers. The file name is made by
3781 appending @file{.ch} to the source file name.
3784 @opindex fdump-tree-ssa
3785 Dump SSA related information to a file. The file name is made by appending
3786 @file{.ssa} to the source file name.
3789 @opindex fdump-tree-alias
3790 Dump aliasing information for each function. The file name is made by
3791 appending @file{.alias} to the source file name.
3794 @opindex fdump-tree-ccp
3795 Dump each function after CCP@. The file name is made by appending
3796 @file{.ccp} to the source file name.
3799 @opindex fdump-tree-pre
3800 Dump trees after partial redundancy elimination. The file name is made
3801 by appending @file{.pre} to the source file name.
3804 @opindex fdump-tree-fre
3805 Dump trees after full redundancy elimination. The file name is made
3806 by appending @file{.fre} to the source file name.
3809 @opindex fdump-tree-dce
3810 Dump each function after dead code elimination. The file name is made by
3811 appending @file{.dce} to the source file name.
3814 @opindex fdump-tree-mudflap
3815 Dump each function after adding mudflap instrumentation. The file name is
3816 made by appending @file{.mudflap} to the source file name.
3819 @opindex fdump-tree-sra
3820 Dump each function after performing scalar replacement of aggregates. The
3821 file name is made by appending @file{.sra} to the source file name.
3824 @opindex fdump-tree-dom
3825 Dump each function after applying dominator tree optimizations. The file
3826 name is made by appending @file{.dom} to the source file name.
3829 @opindex fdump-tree-dse
3830 Dump each function after applying dead store elimination. The file
3831 name is made by appending @file{.dse} to the source file name.
3834 @opindex fdump-tree-phiopt
3835 Dump each function after optimizing PHI nodes into straightline code. The file
3836 name is made by appending @file{.phiopt} to the source file name.
3839 @opindex fdump-tree-forwprop
3840 Dump each function after forward propagating single use variables. The file
3841 name is made by appending @file{.forwprop} to the source file name.
3844 @opindex fdump-tree-copyrename
3845 Dump each function after applying the copy rename optimization. The file
3846 name is made by appending @file{.copyrename} to the source file name.
3849 @opindex fdump-tree-nrv
3850 Dump each function after applying the named return value optimization on
3851 generic trees. The file name is made by appending @file{.nrv} to the source
3855 @opindex fdump-tree-vect
3856 Dump each function after applying vectorization of loops. The file name is
3857 made by appending @file{.vect} to the source file name.
3860 @opindex fdump-tree-all
3861 Enable all the available tree dumps with the flags provided in this option.
3864 @item -frandom-seed=@var{string}
3865 @opindex frandom-string
3866 This option provides a seed that GCC uses when it would otherwise use
3867 random numbers. It is used to generate certain symbol names
3868 that have to be different in every compiled file. It is also used to
3869 place unique stamps in coverage data files and the object files that
3870 produce them. You can use the @option{-frandom-seed} option to produce
3871 reproducibly identical object files.
3873 The @var{string} should be different for every file you compile.
3875 @item -fsched-verbose=@var{n}
3876 @opindex fsched-verbose
3877 On targets that use instruction scheduling, this option controls the
3878 amount of debugging output the scheduler prints. This information is
3879 written to standard error, unless @option{-dS} or @option{-dR} is
3880 specified, in which case it is output to the usual dump
3881 listing file, @file{.sched} or @file{.sched2} respectively. However
3882 for @var{n} greater than nine, the output is always printed to standard
3885 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
3886 same information as @option{-dRS}. For @var{n} greater than one, it
3887 also output basic block probabilities, detailed ready list information
3888 and unit/insn info. For @var{n} greater than two, it includes RTL
3889 at abort point, control-flow and regions info. And for @var{n} over
3890 four, @option{-fsched-verbose} also includes dependence info.
3894 Store the usual ``temporary'' intermediate files permanently; place them
3895 in the current directory and name them based on the source file. Thus,
3896 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3897 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3898 preprocessed @file{foo.i} output file even though the compiler now
3899 normally uses an integrated preprocessor.
3901 When used in combination with the @option{-x} command line option,
3902 @option{-save-temps} is sensible enough to avoid over writing an
3903 input source file with the same extension as an intermediate file.
3904 The corresponding intermediate file may be obtained by renaming the
3905 source file before using @option{-save-temps}.
3909 Report the CPU time taken by each subprocess in the compilation
3910 sequence. For C source files, this is the compiler proper and assembler
3911 (plus the linker if linking is done). The output looks like this:
3918 The first number on each line is the ``user time'', that is time spent
3919 executing the program itself. The second number is ``system time'',
3920 time spent executing operating system routines on behalf of the program.
3921 Both numbers are in seconds.
3923 @item -fvar-tracking
3924 @opindex fvar-tracking
3925 Run variable tracking pass. It computes where variables are stored at each
3926 position in code. Better debugging information is then generated
3927 (if the debugging information format supports this information).
3929 It is enabled by default when compiling with optimization (@option{-Os},
3930 @option{-O}, @option{-O2}, ...), debugging information (@option{-g}) and
3931 the debug info format supports it.
3933 @item -print-file-name=@var{library}
3934 @opindex print-file-name
3935 Print the full absolute name of the library file @var{library} that
3936 would be used when linking---and don't do anything else. With this
3937 option, GCC does not compile or link anything; it just prints the
3940 @item -print-multi-directory
3941 @opindex print-multi-directory
3942 Print the directory name corresponding to the multilib selected by any
3943 other switches present in the command line. This directory is supposed
3944 to exist in @env{GCC_EXEC_PREFIX}.
3946 @item -print-multi-lib
3947 @opindex print-multi-lib
3948 Print the mapping from multilib directory names to compiler switches
3949 that enable them. The directory name is separated from the switches by
3950 @samp{;}, and each switch starts with an @samp{@@} instead of the
3951 @samp{-}, without spaces between multiple switches. This is supposed to
3952 ease shell-processing.
3954 @item -print-prog-name=@var{program}
3955 @opindex print-prog-name
3956 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3958 @item -print-libgcc-file-name
3959 @opindex print-libgcc-file-name
3960 Same as @option{-print-file-name=libgcc.a}.
3962 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3963 but you do want to link with @file{libgcc.a}. You can do
3966 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3969 @item -print-search-dirs
3970 @opindex print-search-dirs
3971 Print the name of the configured installation directory and a list of
3972 program and library directories @command{gcc} will search---and don't do anything else.
3974 This is useful when @command{gcc} prints the error message
3975 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3976 To resolve this you either need to put @file{cpp0} and the other compiler
3977 components where @command{gcc} expects to find them, or you can set the environment
3978 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3979 Don't forget the trailing @samp{/}.
3980 @xref{Environment Variables}.
3983 @opindex dumpmachine
3984 Print the compiler's target machine (for example,
3985 @samp{i686-pc-linux-gnu})---and don't do anything else.
3988 @opindex dumpversion
3989 Print the compiler version (for example, @samp{3.0})---and don't do
3994 Print the compiler's built-in specs---and don't do anything else. (This
3995 is used when GCC itself is being built.) @xref{Spec Files}.
3997 @item -feliminate-unused-debug-types
3998 @opindex feliminate-unused-debug-types
3999 Normally, when producing DWARF2 output, GCC will emit debugging
4000 information for all types declared in a compilation
4001 unit, regardless of whether or not they are actually used
4002 in that compilation unit. Sometimes this is useful, such as
4003 if, in the debugger, you want to cast a value to a type that is
4004 not actually used in your program (but is declared). More often,
4005 however, this results in a significant amount of wasted space.
4006 With this option, GCC will avoid producing debug symbol output
4007 for types that are nowhere used in the source file being compiled.
4010 @node Optimize Options
4011 @section Options That Control Optimization
4012 @cindex optimize options
4013 @cindex options, optimization
4015 These options control various sorts of optimizations.
4017 Without any optimization option, the compiler's goal is to reduce the
4018 cost of compilation and to make debugging produce the expected
4019 results. Statements are independent: if you stop the program with a
4020 breakpoint between statements, you can then assign a new value to any
4021 variable or change the program counter to any other statement in the
4022 function and get exactly the results you would expect from the source
4025 Turning on optimization flags makes the compiler attempt to improve
4026 the performance and/or code size at the expense of compilation time
4027 and possibly the ability to debug the program.
4029 The compiler performs optimization based on the knowledge it has of
4030 the program. Optimization levels @option{-O2} and above, in
4031 particular, enable @emph{unit-at-a-time} mode, which allows the
4032 compiler to consider information gained from later functions in
4033 the file when compiling a function. Compiling multiple files at
4034 once to a single output file in @emph{unit-at-a-time} mode allows
4035 the compiler to use information gained from all of the files when
4036 compiling each of them.
4038 Not all optimizations are controlled directly by a flag. Only
4039 optimizations that have a flag are listed.
4046 Optimize. Optimizing compilation takes somewhat more time, and a lot
4047 more memory for a large function.
4049 With @option{-O}, the compiler tries to reduce code size and execution
4050 time, without performing any optimizations that take a great deal of
4053 @option{-O} turns on the following optimization flags:
4054 @gccoptlist{-fdefer-pop @gol
4055 -fdelayed-branch @gol
4056 -fguess-branch-probability @gol
4057 -fcprop-registers @gol
4058 -floop-optimize @gol
4059 -fif-conversion @gol
4060 -fif-conversion2 @gol
4066 -ftree-live_range_split @gol
4068 -ftree-copyrename @gol
4073 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
4074 where doing so does not interfere with debugging.
4078 Optimize even more. GCC performs nearly all supported optimizations
4079 that do not involve a space-speed tradeoff. The compiler does not
4080 perform loop unrolling or function inlining when you specify @option{-O2}.
4081 As compared to @option{-O}, this option increases both compilation time
4082 and the performance of the generated code.
4084 @option{-O2} turns on all optimization flags specified by @option{-O}. It
4085 also turns on the following optimization flags:
4086 @gccoptlist{-fthread-jumps @gol
4088 -foptimize-sibling-calls @gol
4089 -fcse-follow-jumps -fcse-skip-blocks @gol
4090 -fgcse -fgcse-lm @gol
4091 -fexpensive-optimizations @gol
4092 -fstrength-reduce @gol
4093 -frerun-cse-after-loop -frerun-loop-opt @gol
4097 -fschedule-insns -fschedule-insns2 @gol
4098 -fsched-interblock -fsched-spec @gol
4100 -fstrict-aliasing @gol
4101 -fdelete-null-pointer-checks @gol
4102 -freorder-blocks -freorder-functions @gol
4103 -funit-at-a-time @gol
4104 -falign-functions -falign-jumps @gol
4105 -falign-loops -falign-labels @gol
4108 Please note the warning under @option{-fgcse} about
4109 invoking @option{-O2} on programs that use computed gotos.
4113 Optimize yet more. @option{-O3} turns on all optimizations specified by
4114 @option{-O2} and also turns on the @option{-finline-functions},
4115 @option{-funswitch-loops} and @option{-fgcse-after-reload} options.
4119 Do not optimize. This is the default.
4123 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
4124 do not typically increase code size. It also performs further
4125 optimizations designed to reduce code size.
4127 @option{-Os} disables the following optimization flags:
4128 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
4129 -falign-labels -freorder-blocks -freorder-blocks-and-partition -fprefetch-loop-arrays}
4131 If you use multiple @option{-O} options, with or without level numbers,
4132 the last such option is the one that is effective.
4135 Options of the form @option{-f@var{flag}} specify machine-independent
4136 flags. Most flags have both positive and negative forms; the negative
4137 form of @option{-ffoo} would be @option{-fno-foo}. In the table
4138 below, only one of the forms is listed---the one you typically will
4139 use. You can figure out the other form by either removing @samp{no-}
4142 The following options control specific optimizations. They are either
4143 activated by @option{-O} options or are related to ones that are. You
4144 can use the following flags in the rare cases when ``fine-tuning'' of
4145 optimizations to be performed is desired.
4148 @item -fno-default-inline
4149 @opindex fno-default-inline
4150 Do not make member functions inline by default merely because they are
4151 defined inside the class scope (C++ only). Otherwise, when you specify
4152 @w{@option{-O}}, member functions defined inside class scope are compiled
4153 inline by default; i.e., you don't need to add @samp{inline} in front of
4154 the member function name.
4156 @item -fno-defer-pop
4157 @opindex fno-defer-pop
4158 Always pop the arguments to each function call as soon as that function
4159 returns. For machines which must pop arguments after a function call,
4160 the compiler normally lets arguments accumulate on the stack for several
4161 function calls and pops them all at once.
4163 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4167 Force memory operands to be copied into registers before doing
4168 arithmetic on them. This produces better code by making all memory
4169 references potential common subexpressions. When they are not common
4170 subexpressions, instruction combination should eliminate the separate
4173 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4176 @opindex fforce-addr
4177 Force memory address constants to be copied into registers before
4178 doing arithmetic on them. This may produce better code just as
4179 @option{-fforce-mem} may.
4181 @item -fomit-frame-pointer
4182 @opindex fomit-frame-pointer
4183 Don't keep the frame pointer in a register for functions that
4184 don't need one. This avoids the instructions to save, set up and
4185 restore frame pointers; it also makes an extra register available
4186 in many functions. @strong{It also makes debugging impossible on
4189 On some machines, such as the VAX, this flag has no effect, because
4190 the standard calling sequence automatically handles the frame pointer
4191 and nothing is saved by pretending it doesn't exist. The
4192 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
4193 whether a target machine supports this flag. @xref{Registers,,Register
4194 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
4196 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4198 @item -foptimize-sibling-calls
4199 @opindex foptimize-sibling-calls
4200 Optimize sibling and tail recursive calls.
4202 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4206 Don't pay attention to the @code{inline} keyword. Normally this option
4207 is used to keep the compiler from expanding any functions inline.
4208 Note that if you are not optimizing, no functions can be expanded inline.
4210 @item -finline-functions
4211 @opindex finline-functions
4212 Integrate all simple functions into their callers. The compiler
4213 heuristically decides which functions are simple enough to be worth
4214 integrating in this way.
4216 If all calls to a given function are integrated, and the function is
4217 declared @code{static}, then the function is normally not output as
4218 assembler code in its own right.
4220 Enabled at level @option{-O3}.
4222 @item -finline-limit=@var{n}
4223 @opindex finline-limit
4224 By default, GCC limits the size of functions that can be inlined. This flag
4225 allows the control of this limit for functions that are explicitly marked as
4226 inline (i.e., marked with the inline keyword or defined within the class
4227 definition in c++). @var{n} is the size of functions that can be inlined in
4228 number of pseudo instructions (not counting parameter handling). The default
4229 value of @var{n} is 600.
4230 Increasing this value can result in more inlined code at
4231 the cost of compilation time and memory consumption. Decreasing usually makes
4232 the compilation faster and less code will be inlined (which presumably
4233 means slower programs). This option is particularly useful for programs that
4234 use inlining heavily such as those based on recursive templates with C++.
4236 Inlining is actually controlled by a number of parameters, which may be
4237 specified individually by using @option{--param @var{name}=@var{value}}.
4238 The @option{-finline-limit=@var{n}} option sets some of these parameters
4242 @item max-inline-insns-single
4243 is set to @var{n}/2.
4244 @item max-inline-insns-auto
4245 is set to @var{n}/2.
4246 @item min-inline-insns
4247 is set to 130 or @var{n}/4, whichever is smaller.
4248 @item max-inline-insns-rtl
4252 See below for a documentation of the individual
4253 parameters controlling inlining.
4255 @emph{Note:} pseudo instruction represents, in this particular context, an
4256 abstract measurement of function's size. In no way, it represents a count
4257 of assembly instructions and as such its exact meaning might change from one
4258 release to an another.
4260 @item -fkeep-inline-functions
4261 @opindex fkeep-inline-functions
4262 In C, emit @code{static} functions that are declared @code{inline}
4263 into the object file, even if the function has been inlined into all
4264 of its callers. This switch does not affect functions using the
4265 @code{extern inline} extension in GNU C@. In C++, emit any and all
4266 inline functions into the object file.
4268 @item -fkeep-static-consts
4269 @opindex fkeep-static-consts
4270 Emit variables declared @code{static const} when optimization isn't turned
4271 on, even if the variables aren't referenced.
4273 GCC enables this option by default. If you want to force the compiler to
4274 check if the variable was referenced, regardless of whether or not
4275 optimization is turned on, use the @option{-fno-keep-static-consts} option.
4277 @item -fmerge-constants
4278 Attempt to merge identical constants (string constants and floating point
4279 constants) across compilation units.
4281 This option is the default for optimized compilation if the assembler and
4282 linker support it. Use @option{-fno-merge-constants} to inhibit this
4285 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4287 @item -fmerge-all-constants
4288 Attempt to merge identical constants and identical variables.
4290 This option implies @option{-fmerge-constants}. In addition to
4291 @option{-fmerge-constants} this considers e.g.@: even constant initialized
4292 arrays or initialized constant variables with integral or floating point
4293 types. Languages like C or C++ require each non-automatic variable to
4294 have distinct location, so using this option will result in non-conforming
4297 @item -fmodulo-sched
4298 @opindex fmodulo-sched
4299 Perform swing modulo scheduling immediately before the first scheduling
4300 pass. This pass looks at innermost loops and reorders their
4301 instructions by overlapping different iterations.
4303 @item -fno-branch-count-reg
4304 @opindex fno-branch-count-reg
4305 Do not use ``decrement and branch'' instructions on a count register,
4306 but instead generate a sequence of instructions that decrement a
4307 register, compare it against zero, then branch based upon the result.
4308 This option is only meaningful on architectures that support such
4309 instructions, which include x86, PowerPC, IA-64 and S/390.
4311 The default is @option{-fbranch-count-reg}, enabled when
4312 @option{-fstrength-reduce} is enabled.
4314 @item -fno-function-cse
4315 @opindex fno-function-cse
4316 Do not put function addresses in registers; make each instruction that
4317 calls a constant function contain the function's address explicitly.
4319 This option results in less efficient code, but some strange hacks
4320 that alter the assembler output may be confused by the optimizations
4321 performed when this option is not used.
4323 The default is @option{-ffunction-cse}
4325 @item -fno-zero-initialized-in-bss
4326 @opindex fno-zero-initialized-in-bss
4327 If the target supports a BSS section, GCC by default puts variables that
4328 are initialized to zero into BSS@. This can save space in the resulting
4331 This option turns off this behavior because some programs explicitly
4332 rely on variables going to the data section. E.g., so that the
4333 resulting executable can find the beginning of that section and/or make
4334 assumptions based on that.
4336 The default is @option{-fzero-initialized-in-bss}.
4338 @item -fbounds-check
4339 @opindex fbounds-check
4340 For front-ends that support it, generate additional code to check that
4341 indices used to access arrays are within the declared range. This is
4342 currently only supported by the Java and Fortran front-ends, where
4343 this option defaults to true and false respectively.
4345 @item -fmudflap -fmudflapth -fmudflapir
4349 @cindex bounds checking
4351 For front-ends that support it (C and C++), instrument all risky
4352 pointer/array dereferencing operations, some standard library
4353 string/heap functions, and some other associated constructs with
4354 range/validity tests. Modules so instrumented should be immune to
4355 buffer overflows, invalid heap use, and some other classes of C/C++
4356 programming errors. The instrumentation relies on a separate runtime
4357 library (@file{libmudflap}), which will be linked into a program if
4358 @option{-fmudflap} is given at link time. Run-time behavior of the
4359 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
4360 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
4363 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
4364 link if your program is multi-threaded. Use @option{-fmudflapir}, in
4365 addition to @option{-fmudflap} or @option{-fmudflapth}, if
4366 instrumentation should ignore pointer reads. This produces less
4367 instrumentation (and therefore faster execution) and still provides
4368 some protection against outright memory corrupting writes, but allows
4369 erroneously read data to propagate within a program.
4371 @item -fstrength-reduce
4372 @opindex fstrength-reduce
4373 Perform the optimizations of loop strength reduction and
4374 elimination of iteration variables.
4376 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4378 @item -fthread-jumps
4379 @opindex fthread-jumps
4380 Perform optimizations where we check to see if a jump branches to a
4381 location where another comparison subsumed by the first is found. If
4382 so, the first branch is redirected to either the destination of the
4383 second branch or a point immediately following it, depending on whether
4384 the condition is known to be true or false.
4386 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4388 @item -fcse-follow-jumps
4389 @opindex fcse-follow-jumps
4390 In common subexpression elimination, scan through jump instructions
4391 when the target of the jump is not reached by any other path. For
4392 example, when CSE encounters an @code{if} statement with an
4393 @code{else} clause, CSE will follow the jump when the condition
4396 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4398 @item -fcse-skip-blocks
4399 @opindex fcse-skip-blocks
4400 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
4401 follow jumps which conditionally skip over blocks. When CSE
4402 encounters a simple @code{if} statement with no else clause,
4403 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
4404 body of the @code{if}.
4406 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4408 @item -frerun-cse-after-loop
4409 @opindex frerun-cse-after-loop
4410 Re-run common subexpression elimination after loop optimizations has been
4413 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4415 @item -frerun-loop-opt
4416 @opindex frerun-loop-opt
4417 Run the loop optimizer twice.
4419 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4423 Perform a global common subexpression elimination pass.
4424 This pass also performs global constant and copy propagation.
4426 @emph{Note:} When compiling a program using computed gotos, a GCC
4427 extension, you may get better runtime performance if you disable
4428 the global common subexpression elimination pass by adding
4429 @option{-fno-gcse} to the command line.
4431 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4435 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
4436 attempt to move loads which are only killed by stores into themselves. This
4437 allows a loop containing a load/store sequence to be changed to a load outside
4438 the loop, and a copy/store within the loop.
4440 Enabled by default when gcse is enabled.
4444 When @option{-fgcse-sm} is enabled, a store motion pass is run after
4445 global common subexpression elimination. This pass will attempt to move
4446 stores out of loops. When used in conjunction with @option{-fgcse-lm},
4447 loops containing a load/store sequence can be changed to a load before
4448 the loop and a store after the loop.
4450 Not enabled at any optimization level.
4454 When @option{-fgcse-las} is enabled, the global common subexpression
4455 elimination pass eliminates redundant loads that come after stores to the
4456 same memory location (both partial and full redundancies).
4458 Not enabled at any optimization level.
4460 @item -fgcse-after-reload
4461 @opindex fgcse-after-reload
4462 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
4463 pass is performed after reload. The purpose of this pass is to cleanup
4466 @item -floop-optimize
4467 @opindex floop-optimize
4468 Perform loop optimizations: move constant expressions out of loops, simplify
4469 exit test conditions and optionally do strength-reduction as well.
4471 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4473 @item -floop-optimize2
4474 @opindex floop-optimize2
4475 Perform loop optimizations using the new loop optimizer. The optimizations
4476 (loop unrolling, peeling and unswitching, loop invariant motion) are enabled
4479 @item -fcrossjumping
4480 @opindex crossjumping
4481 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
4482 resulting code may or may not perform better than without cross-jumping.
4484 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4486 @item -fif-conversion
4487 @opindex if-conversion
4488 Attempt to transform conditional jumps into branch-less equivalents. This
4489 include use of conditional moves, min, max, set flags and abs instructions, and
4490 some tricks doable by standard arithmetics. The use of conditional execution
4491 on chips where it is available is controlled by @code{if-conversion2}.
4493 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4495 @item -fif-conversion2
4496 @opindex if-conversion2
4497 Use conditional execution (where available) to transform conditional jumps into
4498 branch-less equivalents.
4500 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4502 @item -fdelete-null-pointer-checks
4503 @opindex fdelete-null-pointer-checks
4504 Use global dataflow analysis to identify and eliminate useless checks
4505 for null pointers. The compiler assumes that dereferencing a null
4506 pointer would have halted the program. If a pointer is checked after
4507 it has already been dereferenced, it cannot be null.
4509 In some environments, this assumption is not true, and programs can
4510 safely dereference null pointers. Use
4511 @option{-fno-delete-null-pointer-checks} to disable this optimization
4512 for programs which depend on that behavior.
4514 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4516 @item -fexpensive-optimizations
4517 @opindex fexpensive-optimizations
4518 Perform a number of minor optimizations that are relatively expensive.
4520 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4522 @item -foptimize-register-move
4524 @opindex foptimize-register-move
4526 Attempt to reassign register numbers in move instructions and as
4527 operands of other simple instructions in order to maximize the amount of
4528 register tying. This is especially helpful on machines with two-operand
4531 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
4534 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4536 @item -fdelayed-branch
4537 @opindex fdelayed-branch
4538 If supported for the target machine, attempt to reorder instructions
4539 to exploit instruction slots available after delayed branch
4542 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4544 @item -fschedule-insns
4545 @opindex fschedule-insns
4546 If supported for the target machine, attempt to reorder instructions to
4547 eliminate execution stalls due to required data being unavailable. This
4548 helps machines that have slow floating point or memory load instructions
4549 by allowing other instructions to be issued until the result of the load
4550 or floating point instruction is required.
4552 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4554 @item -fschedule-insns2
4555 @opindex fschedule-insns2
4556 Similar to @option{-fschedule-insns}, but requests an additional pass of
4557 instruction scheduling after register allocation has been done. This is
4558 especially useful on machines with a relatively small number of
4559 registers and where memory load instructions take more than one cycle.
4561 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4563 @item -fno-sched-interblock
4564 @opindex fno-sched-interblock
4565 Don't schedule instructions across basic blocks. This is normally
4566 enabled by default when scheduling before register allocation, i.e.@:
4567 with @option{-fschedule-insns} or at @option{-O2} or higher.
4569 @item -fno-sched-spec
4570 @opindex fno-sched-spec
4571 Don't allow speculative motion of non-load instructions. This is normally
4572 enabled by default when scheduling before register allocation, i.e.@:
4573 with @option{-fschedule-insns} or at @option{-O2} or higher.
4575 @item -fsched-spec-load
4576 @opindex fsched-spec-load
4577 Allow speculative motion of some load instructions. This only makes
4578 sense when scheduling before register allocation, i.e.@: with
4579 @option{-fschedule-insns} or at @option{-O2} or higher.
4581 @item -fsched-spec-load-dangerous
4582 @opindex fsched-spec-load-dangerous
4583 Allow speculative motion of more load instructions. This only makes
4584 sense when scheduling before register allocation, i.e.@: with
4585 @option{-fschedule-insns} or at @option{-O2} or higher.
4587 @item -fsched-stalled-insns=@var{n}
4588 @opindex fsched-stalled-insns
4589 Define how many insns (if any) can be moved prematurely from the queue
4590 of stalled insns into the ready list, during the second scheduling pass.
4592 @item -fsched-stalled-insns-dep=@var{n}
4593 @opindex fsched-stalled-insns-dep
4594 Define how many insn groups (cycles) will be examined for a dependency
4595 on a stalled insn that is candidate for premature removal from the queue
4596 of stalled insns. Has an effect only during the second scheduling pass,
4597 and only if @option{-fsched-stalled-insns} is used and its value is not zero.
4599 @item -fsched2-use-superblocks
4600 @opindex fsched2-use-superblocks
4601 When scheduling after register allocation, do use superblock scheduling
4602 algorithm. Superblock scheduling allows motion across basic block boundaries
4603 resulting on faster schedules. This option is experimental, as not all machine
4604 descriptions used by GCC model the CPU closely enough to avoid unreliable
4605 results from the algorithm.
4607 This only makes sense when scheduling after register allocation, i.e.@: with
4608 @option{-fschedule-insns2} or at @option{-O2} or higher.
4610 @item -fsched2-use-traces
4611 @opindex fsched2-use-traces
4612 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
4613 allocation and additionally perform code duplication in order to increase the
4614 size of superblocks using tracer pass. See @option{-ftracer} for details on
4617 This mode should produce faster but significantly longer programs. Also
4618 without @option{-fbranch-probabilities} the traces constructed may not
4619 match the reality and hurt the performance. This only makes
4620 sense when scheduling after register allocation, i.e.@: with
4621 @option{-fschedule-insns2} or at @option{-O2} or higher.
4623 @item -freschedule-modulo-scheduled-loops
4624 @opindex fscheduling-in-modulo-scheduled-loops
4625 The modulo scheduling comes before the traditional scheduling, if a loop was modulo scheduled
4626 we may want to prevent the later scheduling passes from changing its schedule, we use this
4627 option to control that.
4629 @item -fcaller-saves
4630 @opindex fcaller-saves
4631 Enable values to be allocated in registers that will be clobbered by
4632 function calls, by emitting extra instructions to save and restore the
4633 registers around such calls. Such allocation is done only when it
4634 seems to result in better code than would otherwise be produced.
4636 This option is always enabled by default on certain machines, usually
4637 those which have no call-preserved registers to use instead.
4639 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4642 Perform Partial Redundancy Elimination (PRE) on trees. This flag is
4643 enabled by default at @option{-O2} and @option{-O3}.
4646 Perform Full Redundancy Elimination (FRE) on trees. The difference
4647 between FRE and PRE is that FRE only considers expressions
4648 that are computed on all paths leading to the redundant computation.
4649 This analysis faster than PRE, though it exposes fewer redundancies.
4650 This flag is enabled by default at @option{-O} and higher.
4653 Perform sparse conditional constant propagation (CCP) on trees. This flag
4654 is enabled by default at @option{-O} and higher.
4657 Perform dead code elimination (DCE) on trees. This flag is enabled by
4658 default at @option{-O} and higher.
4660 @item -ftree-dominator-opts
4661 Perform dead code elimination (DCE) on trees. This flag is enabled by
4662 default at @option{-O} and higher.
4665 Perform loop header copying on trees. This is beneficial since it increases
4666 effectiveness of code motion optimizations. It also saves one jump. This flag
4667 is enabled by default at @option{-O} and higher. It is not enabled
4668 for @option{-Os}, since it usually increases code size.
4670 @item -ftree-loop-optimize
4671 Perform loop optimizations on trees. This flag is enabled by default
4672 at @option{-O} and higher.
4674 @item -ftree-loop-linear
4675 Perform linear loop transformations on tree. This flag can improve cache
4676 performance and allow further loop optimizations to take place.
4679 Perform loop invariant motion on trees. This pass moves only invartiants that
4680 would be hard to handle on rtl level (function calls, operations that expand to
4681 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
4682 operands of conditions that are invariant out of the loop, so that we can use
4683 just trivial invariantness analysis in loop unswitching. The pass also includes
4687 Create a canonical counter for number of iterations in the loop for that
4688 determining number of iterations requires complicated analysis. Later
4689 optimizations then may determine the number easily. Useful especially
4690 in connection with unrolling.
4693 Perform induction variable optimizations (strength reduction, induction
4694 variable merging and induction variable elimination) on trees.
4697 Perform scalar replacement of aggregates. This pass replaces structure
4698 references with scalars to prevent committing structures to memory too
4699 early. This flag is enabled by default at @option{-O} and higher.
4701 @item -ftree-copyrename
4702 Perform copy renaming on trees. This pass attempts to rename compiler
4703 temporaries to other variables at copy locations, usually resulting in
4704 variable names which more closely resemble the original variables. This flag
4705 is enabled by default at @option{-O} and higher.
4708 Perform temporary expression replacement during the SSA->normal phase. Single
4709 use/single def temporaries are replaced at their use location with their
4710 defining expression. This results in non-GIMPLE code, but gives the expanders
4711 much more complex trees to work on resulting in better RTL generation. This is
4712 enabled by default at @option{-O} and higher.
4715 Perform live range splitting during the SSA->normal phase. Distinct live
4716 ranges of a variable are split into unique variables, allowing for better
4717 optimization later. This is enabled by default at @option{-O} and higher.
4719 @item -ftree-vectorize
4720 Perform loop vectorization on trees.
4724 Perform tail duplication to enlarge superblock size. This transformation
4725 simplifies the control flow of the function allowing other optimizations to do
4728 @item -funroll-loops
4729 @opindex funroll-loops
4730 Unroll loops whose number of iterations can be determined at compile
4731 time or upon entry to the loop. @option{-funroll-loops} implies both
4732 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
4733 option makes code larger, and may or may not make it run faster.
4735 @item -funroll-all-loops
4736 @opindex funroll-all-loops
4737 Unroll all loops, even if their number of iterations is uncertain when
4738 the loop is entered. This usually makes programs run more slowly.
4739 @option{-funroll-all-loops} implies the same options as
4740 @option{-funroll-loops},
4742 @item -fsplit-ivs-in-unroller
4743 @opindex -fsplit-ivs-in-unroller
4744 Enables expressing of values of induction variables in later iterations
4745 of the unrolled loop using the value in the first iteration. This breaks
4746 long dependency chains, thus improving efficiency of the scheduling passes
4747 (for best results, @option{-fweb} should be used as well).
4749 Combination of @option{-fweb} and CSE is often sufficient to obtain the
4750 same effect. However in cases the loop body is more complicated than
4751 a single basic block, this is not reliable. It also does not work at all
4752 on some of the architectures due to restrictions in the CSE pass.
4754 This optimization is enabled by default.
4756 @item -fvariable-expansion-in-unroller
4757 @opindex -fvariable-expansion-in-unroller
4758 With this option, the compiler will create multiple copies of some
4759 local variables when unrolling a loop which can result in superior code.
4761 @item -fprefetch-loop-arrays
4762 @opindex fprefetch-loop-arrays
4763 If supported by the target machine, generate instructions to prefetch
4764 memory to improve the performance of loops that access large arrays.
4766 These options may generate better or worse code; results are highly
4767 dependent on the structure of loops within the source code.
4770 @itemx -fno-peephole2
4771 @opindex fno-peephole
4772 @opindex fno-peephole2
4773 Disable any machine-specific peephole optimizations. The difference
4774 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
4775 are implemented in the compiler; some targets use one, some use the
4776 other, a few use both.
4778 @option{-fpeephole} is enabled by default.
4779 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4781 @item -fno-guess-branch-probability
4782 @opindex fno-guess-branch-probability
4783 Do not guess branch probabilities using heuristics.
4785 GCC will use heuristics to guess branch probabilities if they are
4786 not provided by profiling feedback (@option{-fprofile-arcs}). These
4787 heuristics are based on the control flow graph. If some branch probabilities
4788 are specified by @samp{__builtin_expect}, then the heuristics will be
4789 used to guess branch probabilities for the rest of the control flow graph,
4790 taking the @samp{__builtin_expect} info into account. The interactions
4791 between the heuristics and @samp{__builtin_expect} can be complex, and in
4792 some cases, it may be useful to disable the heuristics so that the effects
4793 of @samp{__builtin_expect} are easier to understand.
4795 The default is @option{-fguess-branch-probability} at levels
4796 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4798 @item -freorder-blocks
4799 @opindex freorder-blocks
4800 Reorder basic blocks in the compiled function in order to reduce number of
4801 taken branches and improve code locality.
4803 Enabled at levels @option{-O2}, @option{-O3}.
4805 @item -freorder-blocks-and-partition
4806 @opindex freorder-blocks-and-partition
4807 In addition to reordering basic blocks in the compiled function, in order
4808 to reduce number of taken branches, partitions hot and cold basic blocks
4809 into separate sections of the assembly and .o files, to improve
4810 paging and cache locality performance.
4812 This optimization is automatically turned off in the presence of
4813 exception handling, for linkonce sections, for functions with a user-defined
4814 section attribute and on any architecture that does not support named
4817 @item -freorder-functions
4818 @opindex freorder-functions
4819 Reorder basic blocks in the compiled function in order to reduce number of
4820 taken branches and improve code locality. This is implemented by using special
4821 subsections @code{.text.hot} for most frequently executed functions and
4822 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
4823 the linker so object file format must support named sections and linker must
4824 place them in a reasonable way.
4826 Also profile feedback must be available in to make this option effective. See
4827 @option{-fprofile-arcs} for details.
4829 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4831 @item -fstrict-aliasing
4832 @opindex fstrict-aliasing
4833 Allows the compiler to assume the strictest aliasing rules applicable to
4834 the language being compiled. For C (and C++), this activates
4835 optimizations based on the type of expressions. In particular, an
4836 object of one type is assumed never to reside at the same address as an
4837 object of a different type, unless the types are almost the same. For
4838 example, an @code{unsigned int} can alias an @code{int}, but not a
4839 @code{void*} or a @code{double}. A character type may alias any other
4842 Pay special attention to code like this:
4855 The practice of reading from a different union member than the one most
4856 recently written to (called ``type-punning'') is common. Even with
4857 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
4858 is accessed through the union type. So, the code above will work as
4859 expected. However, this code might not:
4870 Every language that wishes to perform language-specific alias analysis
4871 should define a function that computes, given an @code{tree}
4872 node, an alias set for the node. Nodes in different alias sets are not
4873 allowed to alias. For an example, see the C front-end function
4874 @code{c_get_alias_set}.
4876 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4878 @item -falign-functions
4879 @itemx -falign-functions=@var{n}
4880 @opindex falign-functions
4881 Align the start of functions to the next power-of-two greater than
4882 @var{n}, skipping up to @var{n} bytes. For instance,
4883 @option{-falign-functions=32} aligns functions to the next 32-byte
4884 boundary, but @option{-falign-functions=24} would align to the next
4885 32-byte boundary only if this can be done by skipping 23 bytes or less.
4887 @option{-fno-align-functions} and @option{-falign-functions=1} are
4888 equivalent and mean that functions will not be aligned.
4890 Some assemblers only support this flag when @var{n} is a power of two;
4891 in that case, it is rounded up.
4893 If @var{n} is not specified or is zero, use a machine-dependent default.
4895 Enabled at levels @option{-O2}, @option{-O3}.
4897 @item -falign-labels
4898 @itemx -falign-labels=@var{n}
4899 @opindex falign-labels
4900 Align all branch targets to a power-of-two boundary, skipping up to
4901 @var{n} bytes like @option{-falign-functions}. This option can easily
4902 make code slower, because it must insert dummy operations for when the
4903 branch target is reached in the usual flow of the code.
4905 @option{-fno-align-labels} and @option{-falign-labels=1} are
4906 equivalent and mean that labels will not be aligned.
4908 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
4909 are greater than this value, then their values are used instead.
4911 If @var{n} is not specified or is zero, use a machine-dependent default
4912 which is very likely to be @samp{1}, meaning no alignment.
4914 Enabled at levels @option{-O2}, @option{-O3}.
4917 @itemx -falign-loops=@var{n}
4918 @opindex falign-loops
4919 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
4920 like @option{-falign-functions}. The hope is that the loop will be
4921 executed many times, which will make up for any execution of the dummy
4924 @option{-fno-align-loops} and @option{-falign-loops=1} are
4925 equivalent and mean that loops will not be aligned.
4927 If @var{n} is not specified or is zero, use a machine-dependent default.
4929 Enabled at levels @option{-O2}, @option{-O3}.
4932 @itemx -falign-jumps=@var{n}
4933 @opindex falign-jumps
4934 Align branch targets to a power-of-two boundary, for branch targets
4935 where the targets can only be reached by jumping, skipping up to @var{n}
4936 bytes like @option{-falign-functions}. In this case, no dummy operations
4939 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
4940 equivalent and mean that loops will not be aligned.
4942 If @var{n} is not specified or is zero, use a machine-dependent default.
4944 Enabled at levels @option{-O2}, @option{-O3}.
4946 @item -funit-at-a-time
4947 @opindex funit-at-a-time
4948 Parse the whole compilation unit before starting to produce code.
4949 This allows some extra optimizations to take place but consumes
4950 more memory (in general). There are some compatibility issues
4951 with @emph{unit-at-at-time} mode:
4954 enabling @emph{unit-at-a-time} mode may change the order
4955 in which functions, variables, and top-level @code{asm} statements
4956 are emitted, and will likely break code relying on some particular
4957 ordering. The majority of such top-level @code{asm} statements,
4958 though, can be replaced by @code{section} attributes.
4961 @emph{unit-at-a-time} mode removes unreferenced static variables
4962 and functions are removed. This may result in undefined references
4963 when an @code{asm} statement refers directly to variables or functions
4964 that are otherwise unused. In that case either the variable/function
4965 shall be listed as an operand of the @code{asm} statement operand or,
4966 in the case of top-level @code{asm} statements the attribute @code{used}
4967 shall be used on the declaration.
4970 Static functions now can use non-standard passing conventions that
4971 may break @code{asm} statements calling functions directly. Again,
4972 attribute @code{used} will prevent this behavior.
4975 As a temporary workaround, @option{-fno-unit-at-a-time} can be used,
4976 but this scheme may not be supported by future releases of GCC@.
4978 Enabled at levels @option{-O2}, @option{-O3}.
4982 Constructs webs as commonly used for register allocation purposes and assign
4983 each web individual pseudo register. This allows the register allocation pass
4984 to operate on pseudos directly, but also strengthens several other optimization
4985 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
4986 however, make debugging impossible, since variables will no longer stay in a
4989 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os},
4990 on targets where the default format for debugging information supports
4993 @item -fno-cprop-registers
4994 @opindex fno-cprop-registers
4995 After register allocation and post-register allocation instruction splitting,
4996 we perform a copy-propagation pass to try to reduce scheduling dependencies
4997 and occasionally eliminate the copy.
4999 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5001 @item -fprofile-generate
5002 @opindex fprofile-generate
5004 Enable options usually used for instrumenting application to produce
5005 profile useful for later recompilation with profile feedback based
5006 optimization. You must use @option{-fprofile-generate} both when
5007 compiling and when linking your program.
5009 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
5012 @opindex fprofile-use
5013 Enable profile feedback directed optimizations, and optimizations
5014 generally profitable only with profile feedback available.
5016 The following options are enabled: @code{-fbranch-probabilities},
5017 @code{-fvpt}, @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}.
5021 The following options control compiler behavior regarding floating
5022 point arithmetic. These options trade off between speed and
5023 correctness. All must be specifically enabled.
5027 @opindex ffloat-store
5028 Do not store floating point variables in registers, and inhibit other
5029 options that might change whether a floating point value is taken from a
5032 @cindex floating point precision
5033 This option prevents undesirable excess precision on machines such as
5034 the 68000 where the floating registers (of the 68881) keep more
5035 precision than a @code{double} is supposed to have. Similarly for the
5036 x86 architecture. For most programs, the excess precision does only
5037 good, but a few programs rely on the precise definition of IEEE floating
5038 point. Use @option{-ffloat-store} for such programs, after modifying
5039 them to store all pertinent intermediate computations into variables.
5043 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
5044 @option{-fno-trapping-math}, @option{-ffinite-math-only},
5045 @option{-fno-rounding-math} and @option{-fno-signaling-nans}.
5047 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
5049 This option should never be turned on by any @option{-O} option since
5050 it can result in incorrect output for programs which depend on
5051 an exact implementation of IEEE or ISO rules/specifications for
5054 @item -fno-math-errno
5055 @opindex fno-math-errno
5056 Do not set ERRNO after calling math functions that are executed
5057 with a single instruction, e.g., sqrt. A program that relies on
5058 IEEE exceptions for math error handling may want to use this flag
5059 for speed while maintaining IEEE arithmetic compatibility.
5061 This option should never be turned on by any @option{-O} option since
5062 it can result in incorrect output for programs which depend on
5063 an exact implementation of IEEE or ISO rules/specifications for
5066 The default is @option{-fmath-errno}.
5068 @item -funsafe-math-optimizations
5069 @opindex funsafe-math-optimizations
5070 Allow optimizations for floating-point arithmetic that (a) assume
5071 that arguments and results are valid and (b) may violate IEEE or
5072 ANSI standards. When used at link-time, it may include libraries
5073 or startup files that change the default FPU control word or other
5074 similar optimizations.
5076 This option should never be turned on by any @option{-O} option since
5077 it can result in incorrect output for programs which depend on
5078 an exact implementation of IEEE or ISO rules/specifications for
5081 The default is @option{-fno-unsafe-math-optimizations}.
5083 @item -ffinite-math-only
5084 @opindex ffinite-math-only
5085 Allow optimizations for floating-point arithmetic that assume
5086 that arguments and results are not NaNs or +-Infs.
5088 This option should never be turned on by any @option{-O} option since
5089 it can result in incorrect output for programs which depend on
5090 an exact implementation of IEEE or ISO rules/specifications.
5092 The default is @option{-fno-finite-math-only}.
5094 @item -fno-trapping-math
5095 @opindex fno-trapping-math
5096 Compile code assuming that floating-point operations cannot generate
5097 user-visible traps. These traps include division by zero, overflow,
5098 underflow, inexact result and invalid operation. This option implies
5099 @option{-fno-signaling-nans}. Setting this option may allow faster
5100 code if one relies on ``non-stop'' IEEE arithmetic, for example.
5102 This option should never be turned on by any @option{-O} option since
5103 it can result in incorrect output for programs which depend on
5104 an exact implementation of IEEE or ISO rules/specifications for
5107 The default is @option{-ftrapping-math}.
5109 @item -frounding-math
5110 @opindex frounding-math
5111 Disable transformations and optimizations that assume default floating
5112 point rounding behavior. This is round-to-zero for all floating point
5113 to integer conversions, and round-to-nearest for all other arithmetic
5114 truncations. This option should be specified for programs that change
5115 the FP rounding mode dynamically, or that may be executed with a
5116 non-default rounding mode. This option disables constant folding of
5117 floating point expressions at compile-time (which may be affected by
5118 rounding mode) and arithmetic transformations that are unsafe in the
5119 presence of sign-dependent rounding modes.
5121 The default is @option{-fno-rounding-math}.
5123 This option is experimental and does not currently guarantee to
5124 disable all GCC optimizations that are affected by rounding mode.
5125 Future versions of GCC may provide finer control of this setting
5126 using C99's @code{FENV_ACCESS} pragma. This command line option
5127 will be used to specify the default state for @code{FENV_ACCESS}.
5129 @item -fsignaling-nans
5130 @opindex fsignaling-nans
5131 Compile code assuming that IEEE signaling NaNs may generate user-visible
5132 traps during floating-point operations. Setting this option disables
5133 optimizations that may change the number of exceptions visible with
5134 signaling NaNs. This option implies @option{-ftrapping-math}.
5136 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
5139 The default is @option{-fno-signaling-nans}.
5141 This option is experimental and does not currently guarantee to
5142 disable all GCC optimizations that affect signaling NaN behavior.
5144 @item -fsingle-precision-constant
5145 @opindex fsingle-precision-constant
5146 Treat floating point constant as single precision constant instead of
5147 implicitly converting it to double precision constant.
5152 The following options control optimizations that may improve
5153 performance, but are not enabled by any @option{-O} options. This
5154 section includes experimental options that may produce broken code.
5157 @item -fbranch-probabilities
5158 @opindex fbranch-probabilities
5159 After running a program compiled with @option{-fprofile-arcs}
5160 (@pxref{Debugging Options,, Options for Debugging Your Program or
5161 @command{gcc}}), you can compile it a second time using
5162 @option{-fbranch-probabilities}, to improve optimizations based on
5163 the number of times each branch was taken. When the program
5164 compiled with @option{-fprofile-arcs} exits it saves arc execution
5165 counts to a file called @file{@var{sourcename}.gcda} for each source
5166 file The information in this data file is very dependent on the
5167 structure of the generated code, so you must use the same source code
5168 and the same optimization options for both compilations.
5170 With @option{-fbranch-probabilities}, GCC puts a
5171 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
5172 These can be used to improve optimization. Currently, they are only
5173 used in one place: in @file{reorg.c}, instead of guessing which path a
5174 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
5175 exactly determine which path is taken more often.
5177 @item -fprofile-values
5178 @opindex fprofile-values
5179 If combined with @option{-fprofile-arcs}, it adds code so that some
5180 data about values of expressions in the program is gathered.
5182 With @option{-fbranch-probabilities}, it reads back the data gathered
5183 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
5184 notes to instructions for their later usage in optimizations.
5186 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
5190 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
5191 a code to gather information about values of expressions.
5193 With @option{-fbranch-probabilities}, it reads back the data gathered
5194 and actually performs the optimizations based on them.
5195 Currently the optimizations include specialization of division operation
5196 using the knowledge about the value of the denominator.
5198 @item -fspeculative-prefetching
5199 @opindex fspeculative-prefetching
5200 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
5201 a code to gather information about addresses of memory references in the
5204 With @option{-fbranch-probabilities}, it reads back the data gathered
5205 and issues prefetch instructions according to them. In addition to the opportunities
5206 noticed by @option{-fprefetch-loop-arrays}, it also notices more complicated
5207 memory access patterns---for example accesses to the data stored in linked
5208 list whose elements are usually allocated sequentially.
5210 In order to prevent issuing double prefetches, usage of
5211 @option{-fspeculative-prefetching} implies @option{-fno-prefetch-loop-arrays}.
5213 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
5215 @item -frename-registers
5216 @opindex frename-registers
5217 Attempt to avoid false dependencies in scheduled code by making use
5218 of registers left over after register allocation. This optimization
5219 will most benefit processors with lots of registers. Depending on the
5220 debug information format adopted by the target, however, it can
5221 make debugging impossible, since variables will no longer stay in
5222 a ``home register''.
5224 Not enabled by default at any level because it has known bugs.
5228 Perform tail duplication to enlarge superblock size. This transformation
5229 simplifies the control flow of the function allowing other optimizations to do
5232 Enabled with @option{-fprofile-use}.
5234 @item -funroll-loops
5235 @opindex funroll-loops
5236 Unroll loops whose number of iterations can be determined at compile time or
5237 upon entry to the loop. @option{-funroll-loops} implies
5238 @option{-frerun-cse-after-loop}. It also turns on complete loop peeling
5239 (i.e.@: complete removal of loops with small constant number of iterations).
5240 This option makes code larger, and may or may not make it run faster.
5242 Enabled with @option{-fprofile-use}.
5244 @item -funroll-all-loops
5245 @opindex funroll-all-loops
5246 Unroll all loops, even if their number of iterations is uncertain when
5247 the loop is entered. This usually makes programs run more slowly.
5248 @option{-funroll-all-loops} implies the same options as
5249 @option{-funroll-loops}.
5252 @opindex fpeel-loops
5253 Peels the loops for that there is enough information that they do not
5254 roll much (from profile feedback). It also turns on complete loop peeling
5255 (i.e.@: complete removal of loops with small constant number of iterations).
5257 Enabled with @option{-fprofile-use}.
5259 @item -fmove-loop-invariants
5260 @opindex fmove-loop-invariants
5261 Enables the loop invariant motion pass in the new loop optimizer. Enabled
5262 at level @option{-O1}
5264 @item -funswitch-loops
5265 @opindex funswitch-loops
5266 Move branches with loop invariant conditions out of the loop, with duplicates
5267 of the loop on both branches (modified according to result of the condition).
5269 @item -fprefetch-loop-arrays
5270 @opindex fprefetch-loop-arrays
5271 If supported by the target machine, generate instructions to prefetch
5272 memory to improve the performance of loops that access large arrays.
5274 Disabled at level @option{-Os}.
5276 @item -ffunction-sections
5277 @itemx -fdata-sections
5278 @opindex ffunction-sections
5279 @opindex fdata-sections
5280 Place each function or data item into its own section in the output
5281 file if the target supports arbitrary sections. The name of the
5282 function or the name of the data item determines the section's name
5285 Use these options on systems where the linker can perform optimizations
5286 to improve locality of reference in the instruction space. Most systems
5287 using the ELF object format and SPARC processors running Solaris 2 have
5288 linkers with such optimizations. AIX may have these optimizations in
5291 Only use these options when there are significant benefits from doing
5292 so. When you specify these options, the assembler and linker will
5293 create larger object and executable files and will also be slower.
5294 You will not be able to use @code{gprof} on all systems if you
5295 specify this option and you may have problems with debugging if
5296 you specify both this option and @option{-g}.
5298 @item -fbranch-target-load-optimize
5299 @opindex fbranch-target-load-optimize
5300 Perform branch target register load optimization before prologue / epilogue
5302 The use of target registers can typically be exposed only during reload,
5303 thus hoisting loads out of loops and doing inter-block scheduling needs
5304 a separate optimization pass.
5306 @item -fbranch-target-load-optimize2
5307 @opindex fbranch-target-load-optimize2
5308 Perform branch target register load optimization after prologue / epilogue
5311 @item -fbtr-bb-exclusive
5312 @opindex fbtr-bb-exclusive
5313 When performing branch target register load optimization, don't reuse
5314 branch target registers in within any basic block.
5316 @item --param @var{name}=@var{value}
5318 In some places, GCC uses various constants to control the amount of
5319 optimization that is done. For example, GCC will not inline functions
5320 that contain more that a certain number of instructions. You can
5321 control some of these constants on the command-line using the
5322 @option{--param} option.
5324 The names of specific parameters, and the meaning of the values, are
5325 tied to the internals of the compiler, and are subject to change
5326 without notice in future releases.
5328 In each case, the @var{value} is an integer. The allowable choices for
5329 @var{name} are given in the following table:
5332 @item sra-max-structure-size
5333 The maximum structure size, in bytes, at which the scalar replacement
5334 of aggregates (SRA) optimization will perform block copies. The
5335 default value, 0, implies that GCC will select the most appropriate
5338 @item sra-field-structure-ratio
5339 The threshold ratio (as a percentage) between instantiated fields and
5340 the complete structure size. We say that if the ratio of the number
5341 of bytes in instantiated fields to the number of bytes in the complete
5342 structure exceeds this parameter, then block copies are not used. The
5345 @item max-crossjump-edges
5346 The maximum number of incoming edges to consider for crossjumping.
5347 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
5348 the number of edges incoming to each block. Increasing values mean
5349 more aggressive optimization, making the compile time increase with
5350 probably small improvement in executable size.
5352 @item min-crossjump-insns
5353 The minimum number of instructions which must be matched at the end
5354 of two blocks before crossjumping will be performed on them. This
5355 value is ignored in the case where all instructions in the block being
5356 crossjumped from are matched. The default value is 5.
5358 @item max-goto-duplication-insns
5359 The maximum number of instructions to duplicate to a block that jumps
5360 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
5361 passes, GCC factors computed gotos early in the compilation process,
5362 and unfactors them as late as possible. Only computed jumps at the
5363 end of a basic blocks with no more than max-goto-duplication-insns are
5364 unfactored. The default value is 8.
5366 @item max-delay-slot-insn-search
5367 The maximum number of instructions to consider when looking for an
5368 instruction to fill a delay slot. If more than this arbitrary number of
5369 instructions is searched, the time savings from filling the delay slot
5370 will be minimal so stop searching. Increasing values mean more
5371 aggressive optimization, making the compile time increase with probably
5372 small improvement in executable run time.
5374 @item max-delay-slot-live-search
5375 When trying to fill delay slots, the maximum number of instructions to
5376 consider when searching for a block with valid live register
5377 information. Increasing this arbitrarily chosen value means more
5378 aggressive optimization, increasing the compile time. This parameter
5379 should be removed when the delay slot code is rewritten to maintain the
5382 @item max-gcse-memory
5383 The approximate maximum amount of memory that will be allocated in
5384 order to perform the global common subexpression elimination
5385 optimization. If more memory than specified is required, the
5386 optimization will not be done.
5388 @item max-gcse-passes
5389 The maximum number of passes of GCSE to run. The default is 1.
5391 @item max-pending-list-length
5392 The maximum number of pending dependencies scheduling will allow
5393 before flushing the current state and starting over. Large functions
5394 with few branches or calls can create excessively large lists which
5395 needlessly consume memory and resources.
5397 @item max-inline-insns-single
5398 Several parameters control the tree inliner used in gcc.
5399 This number sets the maximum number of instructions (counted in GCC's
5400 internal representation) in a single function that the tree inliner
5401 will consider for inlining. This only affects functions declared
5402 inline and methods implemented in a class declaration (C++).
5403 The default value is 500.
5405 @item max-inline-insns-auto
5406 When you use @option{-finline-functions} (included in @option{-O3}),
5407 a lot of functions that would otherwise not be considered for inlining
5408 by the compiler will be investigated. To those functions, a different
5409 (more restrictive) limit compared to functions declared inline can
5411 The default value is 120.
5413 @item large-function-insns
5414 The limit specifying really large functions. For functions larger than this
5415 limit after inlining inlining is constrained by
5416 @option{--param large-function-growth}. This parameter is useful primarily
5417 to avoid extreme compilation time caused by non-linear algorithms used by the
5419 This parameter is ignored when @option{-funit-at-a-time} is not used.
5420 The default value is 3000.
5422 @item large-function-growth
5423 Specifies maximal growth of large function caused by inlining in percents.
5424 This parameter is ignored when @option{-funit-at-a-time} is not used.
5425 The default value is 100 which limits large function growth to 2.0 times
5428 @item inline-unit-growth
5429 Specifies maximal overall growth of the compilation unit caused by inlining.
5430 This parameter is ignored when @option{-funit-at-a-time} is not used.
5431 The default value is 50 which limits unit growth to 1.5 times the original
5434 @item max-inline-insns-recursive
5435 @itemx max-inline-insns-recursive-auto
5436 Specifies maximum number of instructions out-of-line copy of self recursive inline
5437 function can grow into by performing recursive inlining.
5439 For functions declared inline @option{--param max-inline-insns-recursive} is
5440 taken into acount. For function not declared inline, recursive inlining
5441 happens only when @option{-finline-functions} (included in @option{-O3}) is
5442 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
5443 default value is 500.
5445 @item max-inline-recursive-depth
5446 @itemx max-inline-recursive-depth-auto
5447 Specifies maximum recursion depth used by the recursive inlining.
5449 For functions declared inline @option{--param max-inline-recursive-depth} is
5450 taken into acount. For function not declared inline, recursive inlining
5451 happens only when @option{-finline-functions} (included in @option{-O3}) is
5452 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
5453 default value is 500.
5455 @item max-unrolled-insns
5456 The maximum number of instructions that a loop should have if that loop
5457 is unrolled, and if the loop is unrolled, it determines how many times
5458 the loop code is unrolled.
5460 @item max-average-unrolled-insns
5461 The maximum number of instructions biased by probabilities of their execution
5462 that a loop should have if that loop is unrolled, and if the loop is unrolled,
5463 it determines how many times the loop code is unrolled.
5465 @item max-unroll-times
5466 The maximum number of unrollings of a single loop.
5468 @item max-peeled-insns
5469 The maximum number of instructions that a loop should have if that loop
5470 is peeled, and if the loop is peeled, it determines how many times
5471 the loop code is peeled.
5473 @item max-peel-times
5474 The maximum number of peelings of a single loop.
5476 @item max-completely-peeled-insns
5477 The maximum number of insns of a completely peeled loop.
5479 @item max-completely-peel-times
5480 The maximum number of iterations of a loop to be suitable for complete peeling.
5482 @item max-unswitch-insns
5483 The maximum number of insns of an unswitched loop.
5485 @item max-unswitch-level
5486 The maximum number of branches unswitched in a single loop.
5489 The minimum cost of an expensive expression in the loop invariant motion.
5491 @item iv-consider-all-candidates-bound
5492 Bound on number of candidates for induction variables below that
5493 all candidates are considered for each use in induction variable
5494 optimizations. Only the most relevant candidates are considered
5495 if there are more candidates, to avoid quadratic time complexity.
5497 @item iv-max-considered-uses
5498 The induction variable optimizations give up on loops that contain more
5499 induction variable uses.
5501 @item iv-always-prune-cand-set-bound
5502 If number of candidates in the set is smaller than this value,
5503 we always try to remove unnecessary ivs from the set during its
5504 optimization when a new iv is added to the set.
5506 @item max-iterations-to-track
5508 The maximum number of iterations of a loop the brute force algorithm
5509 for analysis of # of iterations of the loop tries to evaluate.
5511 @item hot-bb-count-fraction
5512 Select fraction of the maximal count of repetitions of basic block in program
5513 given basic block needs to have to be considered hot.
5515 @item hot-bb-frequency-fraction
5516 Select fraction of the maximal frequency of executions of basic block in
5517 function given basic block needs to have to be considered hot
5519 @item tracer-dynamic-coverage
5520 @itemx tracer-dynamic-coverage-feedback
5522 This value is used to limit superblock formation once the given percentage of
5523 executed instructions is covered. This limits unnecessary code size
5526 The @option{tracer-dynamic-coverage-feedback} is used only when profile
5527 feedback is available. The real profiles (as opposed to statically estimated
5528 ones) are much less balanced allowing the threshold to be larger value.
5530 @item tracer-max-code-growth
5531 Stop tail duplication once code growth has reached given percentage. This is
5532 rather hokey argument, as most of the duplicates will be eliminated later in
5533 cross jumping, so it may be set to much higher values than is the desired code
5536 @item tracer-min-branch-ratio
5538 Stop reverse growth when the reverse probability of best edge is less than this
5539 threshold (in percent).
5541 @item tracer-min-branch-ratio
5542 @itemx tracer-min-branch-ratio-feedback
5544 Stop forward growth if the best edge do have probability lower than this
5547 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
5548 compilation for profile feedback and one for compilation without. The value
5549 for compilation with profile feedback needs to be more conservative (higher) in
5550 order to make tracer effective.
5552 @item max-cse-path-length
5554 Maximum number of basic blocks on path that cse considers. The default is 10.
5556 @item global-var-threshold
5558 Counts the number of function calls (@var{n}) and the number of
5559 call-clobbered variables (@var{v}). If @var{n}x@var{v} is larger than this limit, a
5560 single artificial variable will be created to represent all the
5561 call-clobbered variables at function call sites. This artificial
5562 variable will then be made to alias every call-clobbered variable.
5563 (done as @code{int * size_t} on the host machine; beware overflow).
5565 @item max-aliased-vops
5567 Maximum number of virtual operands allowed to represent aliases
5568 before triggering the alias grouping heuristic. Alias grouping
5569 reduces compile times and memory consumption needed for aliasing at
5570 the expense of precision loss in alias information.
5572 @item ggc-min-expand
5574 GCC uses a garbage collector to manage its own memory allocation. This
5575 parameter specifies the minimum percentage by which the garbage
5576 collector's heap should be allowed to expand between collections.
5577 Tuning this may improve compilation speed; it has no effect on code
5580 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
5581 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
5582 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
5583 GCC is not able to calculate RAM on a particular platform, the lower
5584 bound of 30% is used. Setting this parameter and
5585 @option{ggc-min-heapsize} to zero causes a full collection to occur at
5586 every opportunity. This is extremely slow, but can be useful for
5589 @item ggc-min-heapsize
5591 Minimum size of the garbage collector's heap before it begins bothering
5592 to collect garbage. The first collection occurs after the heap expands
5593 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
5594 tuning this may improve compilation speed, and has no effect on code
5597 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
5598 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
5599 with a lower bound of 4096 (four megabytes) and an upper bound of
5600 131072 (128 megabytes). If GCC is not able to calculate RAM on a
5601 particular platform, the lower bound is used. Setting this parameter
5602 very large effectively disables garbage collection. Setting this
5603 parameter and @option{ggc-min-expand} to zero causes a full collection
5604 to occur at every opportunity.
5606 @item max-reload-search-insns
5607 The maximum number of instruction reload should look backward for equivalent
5608 register. Increasing values mean more aggressive optimization, making the
5609 compile time increase with probably slightly better performance. The default
5612 @item max-cselib-memory-location
5613 The maximum number of memory locations cselib should take into acount.
5614 Increasing values mean more aggressive optimization, making the compile time
5615 increase with probably slightly better performance. The default value is 500.
5617 @item reorder-blocks-duplicate
5618 @itemx reorder-blocks-duplicate-feedback
5620 Used by basic block reordering pass to decide whether to use unconditional
5621 branch or duplicate the code on its destination. Code is duplicated when its
5622 estimated size is smaller than this value multiplied by the estimated size of
5623 unconditional jump in the hot spots of the program.
5625 The @option{reorder-block-duplicate-feedback} is used only when profile
5626 feedback is available and may be set to higher values than
5627 @option{reorder-block-duplicate} since information about the hot spots is more
5630 @item max-sched-region-blocks
5631 The maximum number of blocks in a region to be considered for
5632 interblock scheduling. The default value is 10.
5634 @item max-sched-region-insns
5635 The maximum number of insns in a region to be considered for
5636 interblock scheduling. The default value is 100.
5638 @item max-last-value-rtl
5640 The maximum size measured as number of RTLs that can be recorded in an expression
5641 in combiner for a pseudo register as last known value of that register. The default
5644 @item integer-share-limit
5645 Small integer constants can use a shared data structure, reducing the
5646 compiler's memory usage and increasing its speed. This sets the maximum
5647 value of a shared integer constant's. The default value is 256.
5652 @node Preprocessor Options
5653 @section Options Controlling the Preprocessor
5654 @cindex preprocessor options
5655 @cindex options, preprocessor
5657 These options control the C preprocessor, which is run on each C source
5658 file before actual compilation.
5660 If you use the @option{-E} option, nothing is done except preprocessing.
5661 Some of these options make sense only together with @option{-E} because
5662 they cause the preprocessor output to be unsuitable for actual
5667 You can use @option{-Wp,@var{option}} to bypass the compiler driver
5668 and pass @var{option} directly through to the preprocessor. If
5669 @var{option} contains commas, it is split into multiple options at the
5670 commas. However, many options are modified, translated or interpreted
5671 by the compiler driver before being passed to the preprocessor, and
5672 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
5673 interface is undocumented and subject to change, so whenever possible
5674 you should avoid using @option{-Wp} and let the driver handle the
5677 @item -Xpreprocessor @var{option}
5678 @opindex preprocessor
5679 Pass @var{option} as an option to the preprocessor. You can use this to
5680 supply system-specific preprocessor options which GCC does not know how to
5683 If you want to pass an option that takes an argument, you must use
5684 @option{-Xpreprocessor} twice, once for the option and once for the argument.
5687 @include cppopts.texi
5689 @node Assembler Options
5690 @section Passing Options to the Assembler
5692 @c prevent bad page break with this line
5693 You can pass options to the assembler.
5696 @item -Wa,@var{option}
5698 Pass @var{option} as an option to the assembler. If @var{option}
5699 contains commas, it is split into multiple options at the commas.
5701 @item -Xassembler @var{option}
5703 Pass @var{option} as an option to the assembler. You can use this to
5704 supply system-specific assembler options which GCC does not know how to
5707 If you want to pass an option that takes an argument, you must use
5708 @option{-Xassembler} twice, once for the option and once for the argument.
5713 @section Options for Linking
5714 @cindex link options
5715 @cindex options, linking
5717 These options come into play when the compiler links object files into
5718 an executable output file. They are meaningless if the compiler is
5719 not doing a link step.
5723 @item @var{object-file-name}
5724 A file name that does not end in a special recognized suffix is
5725 considered to name an object file or library. (Object files are
5726 distinguished from libraries by the linker according to the file
5727 contents.) If linking is done, these object files are used as input
5736 If any of these options is used, then the linker is not run, and
5737 object file names should not be used as arguments. @xref{Overall
5741 @item -l@var{library}
5742 @itemx -l @var{library}
5744 Search the library named @var{library} when linking. (The second
5745 alternative with the library as a separate argument is only for
5746 POSIX compliance and is not recommended.)
5748 It makes a difference where in the command you write this option; the
5749 linker searches and processes libraries and object files in the order they
5750 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
5751 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
5752 to functions in @samp{z}, those functions may not be loaded.
5754 The linker searches a standard list of directories for the library,
5755 which is actually a file named @file{lib@var{library}.a}. The linker
5756 then uses this file as if it had been specified precisely by name.
5758 The directories searched include several standard system directories
5759 plus any that you specify with @option{-L}.
5761 Normally the files found this way are library files---archive files
5762 whose members are object files. The linker handles an archive file by
5763 scanning through it for members which define symbols that have so far
5764 been referenced but not defined. But if the file that is found is an
5765 ordinary object file, it is linked in the usual fashion. The only
5766 difference between using an @option{-l} option and specifying a file name
5767 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
5768 and searches several directories.
5772 You need this special case of the @option{-l} option in order to
5773 link an Objective-C or Objective-C++ program.
5776 @opindex nostartfiles
5777 Do not use the standard system startup files when linking.
5778 The standard system libraries are used normally, unless @option{-nostdlib}
5779 or @option{-nodefaultlibs} is used.
5781 @item -nodefaultlibs
5782 @opindex nodefaultlibs
5783 Do not use the standard system libraries when linking.
5784 Only the libraries you specify will be passed to the linker.
5785 The standard startup files are used normally, unless @option{-nostartfiles}
5786 is used. The compiler may generate calls to @code{memcmp},
5787 @code{memset}, @code{memcpy} and @code{memmove}.
5788 These entries are usually resolved by entries in
5789 libc. These entry points should be supplied through some other
5790 mechanism when this option is specified.
5794 Do not use the standard system startup files or libraries when linking.
5795 No startup files and only the libraries you specify will be passed to
5796 the linker. The compiler may generate calls to @code{memcmp}, @code{memset},
5797 @code{memcpy} and @code{memmove}.
5798 These entries are usually resolved by entries in
5799 libc. These entry points should be supplied through some other
5800 mechanism when this option is specified.
5802 @cindex @option{-lgcc}, use with @option{-nostdlib}
5803 @cindex @option{-nostdlib} and unresolved references
5804 @cindex unresolved references and @option{-nostdlib}
5805 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
5806 @cindex @option{-nodefaultlibs} and unresolved references
5807 @cindex unresolved references and @option{-nodefaultlibs}
5808 One of the standard libraries bypassed by @option{-nostdlib} and
5809 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
5810 that GCC uses to overcome shortcomings of particular machines, or special
5811 needs for some languages.
5812 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
5813 Collection (GCC) Internals},
5814 for more discussion of @file{libgcc.a}.)
5815 In most cases, you need @file{libgcc.a} even when you want to avoid
5816 other standard libraries. In other words, when you specify @option{-nostdlib}
5817 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
5818 This ensures that you have no unresolved references to internal GCC
5819 library subroutines. (For example, @samp{__main}, used to ensure C++
5820 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
5821 GNU Compiler Collection (GCC) Internals}.)
5825 Produce a position independent executable on targets which support it.
5826 For predictable results, you must also specify the same set of options
5827 that were used to generate code (@option{-fpie}, @option{-fPIE},
5828 or model suboptions) when you specify this option.
5832 Remove all symbol table and relocation information from the executable.
5836 On systems that support dynamic linking, this prevents linking with the shared
5837 libraries. On other systems, this option has no effect.
5841 Produce a shared object which can then be linked with other objects to
5842 form an executable. Not all systems support this option. For predictable
5843 results, you must also specify the same set of options that were used to
5844 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
5845 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
5846 needs to build supplementary stub code for constructors to work. On
5847 multi-libbed systems, @samp{gcc -shared} must select the correct support
5848 libraries to link against. Failing to supply the correct flags may lead
5849 to subtle defects. Supplying them in cases where they are not necessary
5852 @item -shared-libgcc
5853 @itemx -static-libgcc
5854 @opindex shared-libgcc
5855 @opindex static-libgcc
5856 On systems that provide @file{libgcc} as a shared library, these options
5857 force the use of either the shared or static version respectively.
5858 If no shared version of @file{libgcc} was built when the compiler was
5859 configured, these options have no effect.
5861 There are several situations in which an application should use the
5862 shared @file{libgcc} instead of the static version. The most common
5863 of these is when the application wishes to throw and catch exceptions
5864 across different shared libraries. In that case, each of the libraries
5865 as well as the application itself should use the shared @file{libgcc}.
5867 Therefore, the G++ and GCJ drivers automatically add
5868 @option{-shared-libgcc} whenever you build a shared library or a main
5869 executable, because C++ and Java programs typically use exceptions, so
5870 this is the right thing to do.
5872 If, instead, you use the GCC driver to create shared libraries, you may
5873 find that they will not always be linked with the shared @file{libgcc}.
5874 If GCC finds, at its configuration time, that you have a non-GNU linker
5875 or a GNU linker that does not support option @option{--eh-frame-hdr},
5876 it will link the shared version of @file{libgcc} into shared libraries
5877 by default. Otherwise, it will take advantage of the linker and optimize
5878 away the linking with the shared version of @file{libgcc}, linking with
5879 the static version of libgcc by default. This allows exceptions to
5880 propagate through such shared libraries, without incurring relocation
5881 costs at library load time.
5883 However, if a library or main executable is supposed to throw or catch
5884 exceptions, you must link it using the G++ or GCJ driver, as appropriate
5885 for the languages used in the program, or using the option
5886 @option{-shared-libgcc}, such that it is linked with the shared
5891 Bind references to global symbols when building a shared object. Warn
5892 about any unresolved references (unless overridden by the link editor
5893 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
5896 @item -Xlinker @var{option}
5898 Pass @var{option} as an option to the linker. You can use this to
5899 supply system-specific linker options which GCC does not know how to
5902 If you want to pass an option that takes an argument, you must use
5903 @option{-Xlinker} twice, once for the option and once for the argument.
5904 For example, to pass @option{-assert definitions}, you must write
5905 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
5906 @option{-Xlinker "-assert definitions"}, because this passes the entire
5907 string as a single argument, which is not what the linker expects.
5909 @item -Wl,@var{option}
5911 Pass @var{option} as an option to the linker. If @var{option} contains
5912 commas, it is split into multiple options at the commas.
5914 @item -u @var{symbol}
5916 Pretend the symbol @var{symbol} is undefined, to force linking of
5917 library modules to define it. You can use @option{-u} multiple times with
5918 different symbols to force loading of additional library modules.
5921 @node Directory Options
5922 @section Options for Directory Search
5923 @cindex directory options
5924 @cindex options, directory search
5927 These options specify directories to search for header files, for
5928 libraries and for parts of the compiler:
5933 Add the directory @var{dir} to the head of the list of directories to be
5934 searched for header files. This can be used to override a system header
5935 file, substituting your own version, since these directories are
5936 searched before the system header file directories. However, you should
5937 not use this option to add directories that contain vendor-supplied
5938 system header files (use @option{-isystem} for that). If you use more than
5939 one @option{-I} option, the directories are scanned in left-to-right
5940 order; the standard system directories come after.
5942 If a standard system include directory, or a directory specified with
5943 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
5944 option will be ignored. The directory will still be searched but as a
5945 system directory at its normal position in the system include chain.
5946 This is to ensure that GCC's procedure to fix buggy system headers and
5947 the ordering for the include_next directive are not inadvertently changed.
5948 If you really need to change the search order for system directories,
5949 use the @option{-nostdinc} and/or @option{-isystem} options.
5951 @item -iquote@var{dir}
5953 Add the directory @var{dir} to the head of the list of directories to
5954 be searched for header files only for the case of @samp{#include
5955 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
5956 otherwise just like @option{-I}.
5960 Add directory @var{dir} to the list of directories to be searched
5963 @item -B@var{prefix}
5965 This option specifies where to find the executables, libraries,
5966 include files, and data files of the compiler itself.
5968 The compiler driver program runs one or more of the subprograms
5969 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
5970 @var{prefix} as a prefix for each program it tries to run, both with and
5971 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
5973 For each subprogram to be run, the compiler driver first tries the
5974 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
5975 was not specified, the driver tries two standard prefixes, which are
5976 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
5977 those results in a file name that is found, the unmodified program
5978 name is searched for using the directories specified in your
5979 @env{PATH} environment variable.
5981 The compiler will check to see if the path provided by the @option{-B}
5982 refers to a directory, and if necessary it will add a directory
5983 separator character at the end of the path.
5985 @option{-B} prefixes that effectively specify directory names also apply
5986 to libraries in the linker, because the compiler translates these
5987 options into @option{-L} options for the linker. They also apply to
5988 includes files in the preprocessor, because the compiler translates these
5989 options into @option{-isystem} options for the preprocessor. In this case,
5990 the compiler appends @samp{include} to the prefix.
5992 The run-time support file @file{libgcc.a} can also be searched for using
5993 the @option{-B} prefix, if needed. If it is not found there, the two
5994 standard prefixes above are tried, and that is all. The file is left
5995 out of the link if it is not found by those means.
5997 Another way to specify a prefix much like the @option{-B} prefix is to use
5998 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
6001 As a special kludge, if the path provided by @option{-B} is
6002 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
6003 9, then it will be replaced by @file{[dir/]include}. This is to help
6004 with boot-strapping the compiler.
6006 @item -specs=@var{file}
6008 Process @var{file} after the compiler reads in the standard @file{specs}
6009 file, in order to override the defaults that the @file{gcc} driver
6010 program uses when determining what switches to pass to @file{cc1},
6011 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
6012 @option{-specs=@var{file}} can be specified on the command line, and they
6013 are processed in order, from left to right.
6017 This option has been deprecated. Please use @option{-iquote} instead for
6018 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
6019 Any directories you specify with @option{-I} options before the @option{-I-}
6020 option are searched only for the case of @samp{#include "@var{file}"};
6021 they are not searched for @samp{#include <@var{file}>}.
6023 If additional directories are specified with @option{-I} options after
6024 the @option{-I-}, these directories are searched for all @samp{#include}
6025 directives. (Ordinarily @emph{all} @option{-I} directories are used
6028 In addition, the @option{-I-} option inhibits the use of the current
6029 directory (where the current input file came from) as the first search
6030 directory for @samp{#include "@var{file}"}. There is no way to
6031 override this effect of @option{-I-}. With @option{-I.} you can specify
6032 searching the directory which was current when the compiler was
6033 invoked. That is not exactly the same as what the preprocessor does
6034 by default, but it is often satisfactory.
6036 @option{-I-} does not inhibit the use of the standard system directories
6037 for header files. Thus, @option{-I-} and @option{-nostdinc} are
6044 @section Specifying subprocesses and the switches to pass to them
6047 @command{gcc} is a driver program. It performs its job by invoking a
6048 sequence of other programs to do the work of compiling, assembling and
6049 linking. GCC interprets its command-line parameters and uses these to
6050 deduce which programs it should invoke, and which command-line options
6051 it ought to place on their command lines. This behavior is controlled
6052 by @dfn{spec strings}. In most cases there is one spec string for each
6053 program that GCC can invoke, but a few programs have multiple spec
6054 strings to control their behavior. The spec strings built into GCC can
6055 be overridden by using the @option{-specs=} command-line switch to specify
6058 @dfn{Spec files} are plaintext files that are used to construct spec
6059 strings. They consist of a sequence of directives separated by blank
6060 lines. The type of directive is determined by the first non-whitespace
6061 character on the line and it can be one of the following:
6064 @item %@var{command}
6065 Issues a @var{command} to the spec file processor. The commands that can
6069 @item %include <@var{file}>
6071 Search for @var{file} and insert its text at the current point in the
6074 @item %include_noerr <@var{file}>
6075 @cindex %include_noerr
6076 Just like @samp{%include}, but do not generate an error message if the include
6077 file cannot be found.
6079 @item %rename @var{old_name} @var{new_name}
6081 Rename the spec string @var{old_name} to @var{new_name}.
6085 @item *[@var{spec_name}]:
6086 This tells the compiler to create, override or delete the named spec
6087 string. All lines after this directive up to the next directive or
6088 blank line are considered to be the text for the spec string. If this
6089 results in an empty string then the spec will be deleted. (Or, if the
6090 spec did not exist, then nothing will happened.) Otherwise, if the spec
6091 does not currently exist a new spec will be created. If the spec does
6092 exist then its contents will be overridden by the text of this
6093 directive, unless the first character of that text is the @samp{+}
6094 character, in which case the text will be appended to the spec.
6096 @item [@var{suffix}]:
6097 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
6098 and up to the next directive or blank line are considered to make up the
6099 spec string for the indicated suffix. When the compiler encounters an
6100 input file with the named suffix, it will processes the spec string in
6101 order to work out how to compile that file. For example:
6108 This says that any input file whose name ends in @samp{.ZZ} should be
6109 passed to the program @samp{z-compile}, which should be invoked with the
6110 command-line switch @option{-input} and with the result of performing the
6111 @samp{%i} substitution. (See below.)
6113 As an alternative to providing a spec string, the text that follows a
6114 suffix directive can be one of the following:
6117 @item @@@var{language}
6118 This says that the suffix is an alias for a known @var{language}. This is
6119 similar to using the @option{-x} command-line switch to GCC to specify a
6120 language explicitly. For example:
6127 Says that .ZZ files are, in fact, C++ source files.
6130 This causes an error messages saying:
6133 @var{name} compiler not installed on this system.
6137 GCC already has an extensive list of suffixes built into it.
6138 This directive will add an entry to the end of the list of suffixes, but
6139 since the list is searched from the end backwards, it is effectively
6140 possible to override earlier entries using this technique.
6144 GCC has the following spec strings built into it. Spec files can
6145 override these strings or create their own. Note that individual
6146 targets can also add their own spec strings to this list.
6149 asm Options to pass to the assembler
6150 asm_final Options to pass to the assembler post-processor
6151 cpp Options to pass to the C preprocessor
6152 cc1 Options to pass to the C compiler
6153 cc1plus Options to pass to the C++ compiler
6154 endfile Object files to include at the end of the link
6155 link Options to pass to the linker
6156 lib Libraries to include on the command line to the linker
6157 libgcc Decides which GCC support library to pass to the linker
6158 linker Sets the name of the linker
6159 predefines Defines to be passed to the C preprocessor
6160 signed_char Defines to pass to CPP to say whether @code{char} is signed
6162 startfile Object files to include at the start of the link
6165 Here is a small example of a spec file:
6171 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
6174 This example renames the spec called @samp{lib} to @samp{old_lib} and
6175 then overrides the previous definition of @samp{lib} with a new one.
6176 The new definition adds in some extra command-line options before
6177 including the text of the old definition.
6179 @dfn{Spec strings} are a list of command-line options to be passed to their
6180 corresponding program. In addition, the spec strings can contain
6181 @samp{%}-prefixed sequences to substitute variable text or to
6182 conditionally insert text into the command line. Using these constructs
6183 it is possible to generate quite complex command lines.
6185 Here is a table of all defined @samp{%}-sequences for spec
6186 strings. Note that spaces are not generated automatically around the
6187 results of expanding these sequences. Therefore you can concatenate them
6188 together or combine them with constant text in a single argument.
6192 Substitute one @samp{%} into the program name or argument.
6195 Substitute the name of the input file being processed.
6198 Substitute the basename of the input file being processed.
6199 This is the substring up to (and not including) the last period
6200 and not including the directory.
6203 This is the same as @samp{%b}, but include the file suffix (text after
6207 Marks the argument containing or following the @samp{%d} as a
6208 temporary file name, so that that file will be deleted if GCC exits
6209 successfully. Unlike @samp{%g}, this contributes no text to the
6212 @item %g@var{suffix}
6213 Substitute a file name that has suffix @var{suffix} and is chosen
6214 once per compilation, and mark the argument in the same way as
6215 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
6216 name is now chosen in a way that is hard to predict even when previously
6217 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
6218 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
6219 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
6220 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
6221 was simply substituted with a file name chosen once per compilation,
6222 without regard to any appended suffix (which was therefore treated
6223 just like ordinary text), making such attacks more likely to succeed.
6225 @item %u@var{suffix}
6226 Like @samp{%g}, but generates a new temporary file name even if
6227 @samp{%u@var{suffix}} was already seen.
6229 @item %U@var{suffix}
6230 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
6231 new one if there is no such last file name. In the absence of any
6232 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
6233 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
6234 would involve the generation of two distinct file names, one
6235 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
6236 simply substituted with a file name chosen for the previous @samp{%u},
6237 without regard to any appended suffix.
6239 @item %j@var{suffix}
6240 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
6241 writable, and if save-temps is off; otherwise, substitute the name
6242 of a temporary file, just like @samp{%u}. This temporary file is not
6243 meant for communication between processes, but rather as a junk
6246 @item %|@var{suffix}
6247 @itemx %m@var{suffix}
6248 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
6249 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
6250 all. These are the two most common ways to instruct a program that it
6251 should read from standard input or write to standard output. If you
6252 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
6253 construct: see for example @file{f/lang-specs.h}.
6255 @item %.@var{SUFFIX}
6256 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
6257 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
6258 terminated by the next space or %.
6261 Marks the argument containing or following the @samp{%w} as the
6262 designated output file of this compilation. This puts the argument
6263 into the sequence of arguments that @samp{%o} will substitute later.
6266 Substitutes the names of all the output files, with spaces
6267 automatically placed around them. You should write spaces
6268 around the @samp{%o} as well or the results are undefined.
6269 @samp{%o} is for use in the specs for running the linker.
6270 Input files whose names have no recognized suffix are not compiled
6271 at all, but they are included among the output files, so they will
6275 Substitutes the suffix for object files. Note that this is
6276 handled specially when it immediately follows @samp{%g, %u, or %U},
6277 because of the need for those to form complete file names. The
6278 handling is such that @samp{%O} is treated exactly as if it had already
6279 been substituted, except that @samp{%g, %u, and %U} do not currently
6280 support additional @var{suffix} characters following @samp{%O} as they would
6281 following, for example, @samp{.o}.
6284 Substitutes the standard macro predefinitions for the
6285 current target machine. Use this when running @code{cpp}.
6288 Like @samp{%p}, but puts @samp{__} before and after the name of each
6289 predefined macro, except for macros that start with @samp{__} or with
6290 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
6294 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
6295 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}), and
6296 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
6300 Current argument is the name of a library or startup file of some sort.
6301 Search for that file in a standard list of directories and substitute
6302 the full name found.
6305 Print @var{str} as an error message. @var{str} is terminated by a newline.
6306 Use this when inconsistent options are detected.
6309 Substitute the contents of spec string @var{name} at this point.
6312 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
6314 @item %x@{@var{option}@}
6315 Accumulate an option for @samp{%X}.
6318 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
6322 Output the accumulated assembler options specified by @option{-Wa}.
6325 Output the accumulated preprocessor options specified by @option{-Wp}.
6328 Process the @code{asm} spec. This is used to compute the
6329 switches to be passed to the assembler.
6332 Process the @code{asm_final} spec. This is a spec string for
6333 passing switches to an assembler post-processor, if such a program is
6337 Process the @code{link} spec. This is the spec for computing the
6338 command line passed to the linker. Typically it will make use of the
6339 @samp{%L %G %S %D and %E} sequences.
6342 Dump out a @option{-L} option for each directory that GCC believes might
6343 contain startup files. If the target supports multilibs then the
6344 current multilib directory will be prepended to each of these paths.
6347 Output the multilib directory with directory separators replaced with
6348 @samp{_}. If multilib directories are not set, or the multilib directory is
6349 @file{.} then this option emits nothing.
6352 Process the @code{lib} spec. This is a spec string for deciding which
6353 libraries should be included on the command line to the linker.
6356 Process the @code{libgcc} spec. This is a spec string for deciding
6357 which GCC support library should be included on the command line to the linker.
6360 Process the @code{startfile} spec. This is a spec for deciding which
6361 object files should be the first ones passed to the linker. Typically
6362 this might be a file named @file{crt0.o}.
6365 Process the @code{endfile} spec. This is a spec string that specifies
6366 the last object files that will be passed to the linker.
6369 Process the @code{cpp} spec. This is used to construct the arguments
6370 to be passed to the C preprocessor.
6373 Process the @code{cc1} spec. This is used to construct the options to be
6374 passed to the actual C compiler (@samp{cc1}).
6377 Process the @code{cc1plus} spec. This is used to construct the options to be
6378 passed to the actual C++ compiler (@samp{cc1plus}).
6381 Substitute the variable part of a matched option. See below.
6382 Note that each comma in the substituted string is replaced by
6386 Remove all occurrences of @code{-S} from the command line. Note---this
6387 command is position dependent. @samp{%} commands in the spec string
6388 before this one will see @code{-S}, @samp{%} commands in the spec string
6389 after this one will not.
6391 @item %:@var{function}(@var{args})
6392 Call the named function @var{function}, passing it @var{args}.
6393 @var{args} is first processed as a nested spec string, then split
6394 into an argument vector in the usual fashion. The function returns
6395 a string which is processed as if it had appeared literally as part
6396 of the current spec.
6398 The following built-in spec functions are provided:
6401 @item @code{if-exists}
6402 The @code{if-exists} spec function takes one argument, an absolute
6403 pathname to a file. If the file exists, @code{if-exists} returns the
6404 pathname. Here is a small example of its usage:
6408 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
6411 @item @code{if-exists-else}
6412 The @code{if-exists-else} spec function is similar to the @code{if-exists}
6413 spec function, except that it takes two arguments. The first argument is
6414 an absolute pathname to a file. If the file exists, @code{if-exists-else}
6415 returns the pathname. If it does not exist, it returns the second argument.
6416 This way, @code{if-exists-else} can be used to select one file or another,
6417 based on the existence of the first. Here is a small example of its usage:
6421 crt0%O%s %:if-exists(crti%O%s) \
6422 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
6425 @item @code{replace-outfile}
6426 The @code{replace-outfile} spec function takes two arguments. It looks for the
6427 first argument in the outfiles array and replaces it with the second argument. Here
6428 is a small example of its usage:
6431 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
6437 Substitutes the @code{-S} switch, if that switch was given to GCC@.
6438 If that switch was not specified, this substitutes nothing. Note that
6439 the leading dash is omitted when specifying this option, and it is
6440 automatically inserted if the substitution is performed. Thus the spec
6441 string @samp{%@{foo@}} would match the command-line option @option{-foo}
6442 and would output the command line option @option{-foo}.
6444 @item %W@{@code{S}@}
6445 Like %@{@code{S}@} but mark last argument supplied within as a file to be
6448 @item %@{@code{S}*@}
6449 Substitutes all the switches specified to GCC whose names start
6450 with @code{-S}, but which also take an argument. This is used for
6451 switches like @option{-o}, @option{-D}, @option{-I}, etc.
6452 GCC considers @option{-o foo} as being
6453 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
6454 text, including the space. Thus two arguments would be generated.
6456 @item %@{@code{S}*&@code{T}*@}
6457 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
6458 (the order of @code{S} and @code{T} in the spec is not significant).
6459 There can be any number of ampersand-separated variables; for each the
6460 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
6462 @item %@{@code{S}:@code{X}@}
6463 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
6465 @item %@{!@code{S}:@code{X}@}
6466 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
6468 @item %@{@code{S}*:@code{X}@}
6469 Substitutes @code{X} if one or more switches whose names start with
6470 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
6471 once, no matter how many such switches appeared. However, if @code{%*}
6472 appears somewhere in @code{X}, then @code{X} will be substituted once
6473 for each matching switch, with the @code{%*} replaced by the part of
6474 that switch that matched the @code{*}.
6476 @item %@{.@code{S}:@code{X}@}
6477 Substitutes @code{X}, if processing a file with suffix @code{S}.
6479 @item %@{!.@code{S}:@code{X}@}
6480 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
6482 @item %@{@code{S}|@code{P}:@code{X}@}
6483 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@.
6484 This may be combined with @samp{!}, @samp{.}, and @code{*} sequences as well,
6485 although they have a stronger binding than the @samp{|}. If @code{%*}
6486 appears in @code{X}, all of the alternatives must be starred, and only
6487 the first matching alternative is substituted.
6489 For example, a spec string like this:
6492 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
6495 will output the following command-line options from the following input
6496 command-line options:
6501 -d fred.c -foo -baz -boggle
6502 -d jim.d -bar -baz -boggle
6505 @item %@{S:X; T:Y; :D@}
6507 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
6508 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
6509 be as many clauses as you need. This may be combined with @code{.},
6510 @code{!}, @code{|}, and @code{*} as needed.
6515 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
6516 construct may contain other nested @samp{%} constructs or spaces, or
6517 even newlines. They are processed as usual, as described above.
6518 Trailing white space in @code{X} is ignored. White space may also
6519 appear anywhere on the left side of the colon in these constructs,
6520 except between @code{.} or @code{*} and the corresponding word.
6522 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
6523 handled specifically in these constructs. If another value of
6524 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
6525 @option{-W} switch is found later in the command line, the earlier
6526 switch value is ignored, except with @{@code{S}*@} where @code{S} is
6527 just one letter, which passes all matching options.
6529 The character @samp{|} at the beginning of the predicate text is used to
6530 indicate that a command should be piped to the following command, but
6531 only if @option{-pipe} is specified.
6533 It is built into GCC which switches take arguments and which do not.
6534 (You might think it would be useful to generalize this to allow each
6535 compiler's spec to say which switches take arguments. But this cannot
6536 be done in a consistent fashion. GCC cannot even decide which input
6537 files have been specified without knowing which switches take arguments,
6538 and it must know which input files to compile in order to tell which
6541 GCC also knows implicitly that arguments starting in @option{-l} are to be
6542 treated as compiler output files, and passed to the linker in their
6543 proper position among the other output files.
6545 @c man begin OPTIONS
6547 @node Target Options
6548 @section Specifying Target Machine and Compiler Version
6549 @cindex target options
6550 @cindex cross compiling
6551 @cindex specifying machine version
6552 @cindex specifying compiler version and target machine
6553 @cindex compiler version, specifying
6554 @cindex target machine, specifying
6556 The usual way to run GCC is to run the executable called @file{gcc}, or
6557 @file{<machine>-gcc} when cross-compiling, or
6558 @file{<machine>-gcc-<version>} to run a version other than the one that
6559 was installed last. Sometimes this is inconvenient, so GCC provides
6560 options that will switch to another cross-compiler or version.
6563 @item -b @var{machine}
6565 The argument @var{machine} specifies the target machine for compilation.
6567 The value to use for @var{machine} is the same as was specified as the
6568 machine type when configuring GCC as a cross-compiler. For
6569 example, if a cross-compiler was configured with @samp{configure
6570 i386v}, meaning to compile for an 80386 running System V, then you
6571 would specify @option{-b i386v} to run that cross compiler.
6573 @item -V @var{version}
6575 The argument @var{version} specifies which version of GCC to run.
6576 This is useful when multiple versions are installed. For example,
6577 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
6580 The @option{-V} and @option{-b} options work by running the
6581 @file{<machine>-gcc-<version>} executable, so there's no real reason to
6582 use them if you can just run that directly.
6584 @node Submodel Options
6585 @section Hardware Models and Configurations
6586 @cindex submodel options
6587 @cindex specifying hardware config
6588 @cindex hardware models and configurations, specifying
6589 @cindex machine dependent options
6591 Earlier we discussed the standard option @option{-b} which chooses among
6592 different installed compilers for completely different target
6593 machines, such as VAX vs.@: 68000 vs.@: 80386.
6595 In addition, each of these target machine types can have its own
6596 special options, starting with @samp{-m}, to choose among various
6597 hardware models or configurations---for example, 68010 vs 68020,
6598 floating coprocessor or none. A single installed version of the
6599 compiler can compile for any model or configuration, according to the
6602 Some configurations of the compiler also support additional special
6603 options, usually for compatibility with other compilers on the same
6606 These options are defined by the macro @code{TARGET_SWITCHES} in the
6607 machine description. The default for the options is also defined by
6608 that macro, which enables you to change the defaults.
6610 @c This list is ordered alphanumerically by subsection name.
6611 @c It should be the same order and spelling as these options are listed
6612 @c in Machine Dependent Options
6620 * DEC Alpha Options::
6621 * DEC Alpha/VMS Options::
6625 * i386 and x86-64 Options::
6637 * RS/6000 and PowerPC Options::
6638 * S/390 and zSeries Options::
6641 * System V Options::
6642 * TMS320C3x/C4x Options::
6646 * Xstormy16 Options::
6652 @subsection ARC Options
6655 These options are defined for ARC implementations:
6660 Compile code for little endian mode. This is the default.
6664 Compile code for big endian mode.
6667 @opindex mmangle-cpu
6668 Prepend the name of the cpu to all public symbol names.
6669 In multiple-processor systems, there are many ARC variants with different
6670 instruction and register set characteristics. This flag prevents code
6671 compiled for one cpu to be linked with code compiled for another.
6672 No facility exists for handling variants that are ``almost identical''.
6673 This is an all or nothing option.
6675 @item -mcpu=@var{cpu}
6677 Compile code for ARC variant @var{cpu}.
6678 Which variants are supported depend on the configuration.
6679 All variants support @option{-mcpu=base}, this is the default.
6681 @item -mtext=@var{text-section}
6682 @itemx -mdata=@var{data-section}
6683 @itemx -mrodata=@var{readonly-data-section}
6687 Put functions, data, and readonly data in @var{text-section},
6688 @var{data-section}, and @var{readonly-data-section} respectively
6689 by default. This can be overridden with the @code{section} attribute.
6690 @xref{Variable Attributes}.
6695 @subsection ARM Options
6698 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
6702 @item -mabi=@var{name}
6704 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
6705 @samp{atpcs}, @samp{aapcs} and @samp{iwmmxt}.
6708 @opindex mapcs-frame
6709 Generate a stack frame that is compliant with the ARM Procedure Call
6710 Standard for all functions, even if this is not strictly necessary for
6711 correct execution of the code. Specifying @option{-fomit-frame-pointer}
6712 with this option will cause the stack frames not to be generated for
6713 leaf functions. The default is @option{-mno-apcs-frame}.
6717 This is a synonym for @option{-mapcs-frame}.
6720 @c not currently implemented
6721 @item -mapcs-stack-check
6722 @opindex mapcs-stack-check
6723 Generate code to check the amount of stack space available upon entry to
6724 every function (that actually uses some stack space). If there is
6725 insufficient space available then either the function
6726 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
6727 called, depending upon the amount of stack space required. The run time
6728 system is required to provide these functions. The default is
6729 @option{-mno-apcs-stack-check}, since this produces smaller code.
6731 @c not currently implemented
6733 @opindex mapcs-float
6734 Pass floating point arguments using the float point registers. This is
6735 one of the variants of the APCS@. This option is recommended if the
6736 target hardware has a floating point unit or if a lot of floating point
6737 arithmetic is going to be performed by the code. The default is
6738 @option{-mno-apcs-float}, since integer only code is slightly increased in
6739 size if @option{-mapcs-float} is used.
6741 @c not currently implemented
6742 @item -mapcs-reentrant
6743 @opindex mapcs-reentrant
6744 Generate reentrant, position independent code. The default is
6745 @option{-mno-apcs-reentrant}.
6748 @item -mthumb-interwork
6749 @opindex mthumb-interwork
6750 Generate code which supports calling between the ARM and Thumb
6751 instruction sets. Without this option the two instruction sets cannot
6752 be reliably used inside one program. The default is
6753 @option{-mno-thumb-interwork}, since slightly larger code is generated
6754 when @option{-mthumb-interwork} is specified.
6756 @item -mno-sched-prolog
6757 @opindex mno-sched-prolog
6758 Prevent the reordering of instructions in the function prolog, or the
6759 merging of those instruction with the instructions in the function's
6760 body. This means that all functions will start with a recognizable set
6761 of instructions (or in fact one of a choice from a small set of
6762 different function prologues), and this information can be used to
6763 locate the start if functions inside an executable piece of code. The
6764 default is @option{-msched-prolog}.
6767 @opindex mhard-float
6768 Generate output containing floating point instructions. This is the
6772 @opindex msoft-float
6773 Generate output containing library calls for floating point.
6774 @strong{Warning:} the requisite libraries are not available for all ARM
6775 targets. Normally the facilities of the machine's usual C compiler are
6776 used, but this cannot be done directly in cross-compilation. You must make
6777 your own arrangements to provide suitable library functions for
6780 @option{-msoft-float} changes the calling convention in the output file;
6781 therefore, it is only useful if you compile @emph{all} of a program with
6782 this option. In particular, you need to compile @file{libgcc.a}, the
6783 library that comes with GCC, with @option{-msoft-float} in order for
6786 @item -mfloat-abi=@var{name}
6788 Specifies which ABI to use for floating point values. Permissible values
6789 are: @samp{soft}, @samp{softfp} and @samp{hard}.
6791 @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float}
6792 and @option{-mhard-float} respectively. @samp{softfp} allows the generation
6793 of floating point instructions, but still uses the soft-float calling
6796 @item -mlittle-endian
6797 @opindex mlittle-endian
6798 Generate code for a processor running in little-endian mode. This is
6799 the default for all standard configurations.
6802 @opindex mbig-endian
6803 Generate code for a processor running in big-endian mode; the default is
6804 to compile code for a little-endian processor.
6806 @item -mwords-little-endian
6807 @opindex mwords-little-endian
6808 This option only applies when generating code for big-endian processors.
6809 Generate code for a little-endian word order but a big-endian byte
6810 order. That is, a byte order of the form @samp{32107654}. Note: this
6811 option should only be used if you require compatibility with code for
6812 big-endian ARM processors generated by versions of the compiler prior to
6815 @item -mcpu=@var{name}
6817 This specifies the name of the target ARM processor. GCC uses this name
6818 to determine what kind of instructions it can emit when generating
6819 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
6820 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
6821 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
6822 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
6823 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
6824 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
6825 @samp{arm8}, @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
6826 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
6827 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
6828 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
6829 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
6830 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
6831 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
6832 @samp{arm1176jz-s}, @samp{arm1176jzf-s}, @samp{xscale}, @samp{iwmmxt},
6835 @itemx -mtune=@var{name}
6837 This option is very similar to the @option{-mcpu=} option, except that
6838 instead of specifying the actual target processor type, and hence
6839 restricting which instructions can be used, it specifies that GCC should
6840 tune the performance of the code as if the target were of the type
6841 specified in this option, but still choosing the instructions that it
6842 will generate based on the cpu specified by a @option{-mcpu=} option.
6843 For some ARM implementations better performance can be obtained by using
6846 @item -march=@var{name}
6848 This specifies the name of the target ARM architecture. GCC uses this
6849 name to determine what kind of instructions it can emit when generating
6850 assembly code. This option can be used in conjunction with or instead
6851 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6852 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6853 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j},
6854 @samp{iwmmxt}, @samp{ep9312}.
6856 @item -mfpu=@var{name}
6857 @itemx -mfpe=@var{number}
6858 @itemx -mfp=@var{number}
6862 This specifies what floating point hardware (or hardware emulation) is
6863 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
6864 @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe}
6865 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
6866 with older versions of GCC@.
6868 If @option{-msoft-float} is specified this specifies the format of
6869 floating point values.
6871 @item -mstructure-size-boundary=@var{n}
6872 @opindex mstructure-size-boundary
6873 The size of all structures and unions will be rounded up to a multiple
6874 of the number of bits set by this option. Permissible values are 8, 32
6875 and 64. The default value varies for different toolchains. For the COFF
6876 targeted toolchain the default value is 8. A value of 64 is only allowed
6877 if the underlying ABI supports it.
6879 Specifying the larger number can produce faster, more efficient code, but
6880 can also increase the size of the program. Different values are potentially
6881 incompatible. Code compiled with one value cannot necessarily expect to
6882 work with code or libraries compiled with another value, if they exchange
6883 information using structures or unions.
6885 @item -mabort-on-noreturn
6886 @opindex mabort-on-noreturn
6887 Generate a call to the function @code{abort} at the end of a
6888 @code{noreturn} function. It will be executed if the function tries to
6892 @itemx -mno-long-calls
6893 @opindex mlong-calls
6894 @opindex mno-long-calls
6895 Tells the compiler to perform function calls by first loading the
6896 address of the function into a register and then performing a subroutine
6897 call on this register. This switch is needed if the target function
6898 will lie outside of the 64 megabyte addressing range of the offset based
6899 version of subroutine call instruction.
6901 Even if this switch is enabled, not all function calls will be turned
6902 into long calls. The heuristic is that static functions, functions
6903 which have the @samp{short-call} attribute, functions that are inside
6904 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6905 definitions have already been compiled within the current compilation
6906 unit, will not be turned into long calls. The exception to this rule is
6907 that weak function definitions, functions with the @samp{long-call}
6908 attribute or the @samp{section} attribute, and functions that are within
6909 the scope of a @samp{#pragma long_calls} directive, will always be
6910 turned into long calls.
6912 This feature is not enabled by default. Specifying
6913 @option{-mno-long-calls} will restore the default behavior, as will
6914 placing the function calls within the scope of a @samp{#pragma
6915 long_calls_off} directive. Note these switches have no effect on how
6916 the compiler generates code to handle function calls via function
6919 @item -mnop-fun-dllimport
6920 @opindex mnop-fun-dllimport
6921 Disable support for the @code{dllimport} attribute.
6923 @item -msingle-pic-base
6924 @opindex msingle-pic-base
6925 Treat the register used for PIC addressing as read-only, rather than
6926 loading it in the prologue for each function. The run-time system is
6927 responsible for initializing this register with an appropriate value
6928 before execution begins.
6930 @item -mpic-register=@var{reg}
6931 @opindex mpic-register
6932 Specify the register to be used for PIC addressing. The default is R10
6933 unless stack-checking is enabled, when R9 is used.
6935 @item -mcirrus-fix-invalid-insns
6936 @opindex mcirrus-fix-invalid-insns
6937 @opindex mno-cirrus-fix-invalid-insns
6938 Insert NOPs into the instruction stream to in order to work around
6939 problems with invalid Maverick instruction combinations. This option
6940 is only valid if the @option{-mcpu=ep9312} option has been used to
6941 enable generation of instructions for the Cirrus Maverick floating
6942 point co-processor. This option is not enabled by default, since the
6943 problem is only present in older Maverick implementations. The default
6944 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
6947 @item -mpoke-function-name
6948 @opindex mpoke-function-name
6949 Write the name of each function into the text section, directly
6950 preceding the function prologue. The generated code is similar to this:
6954 .ascii "arm_poke_function_name", 0
6957 .word 0xff000000 + (t1 - t0)
6958 arm_poke_function_name
6960 stmfd sp!, @{fp, ip, lr, pc@}
6964 When performing a stack backtrace, code can inspect the value of
6965 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6966 location @code{pc - 12} and the top 8 bits are set, then we know that
6967 there is a function name embedded immediately preceding this location
6968 and has length @code{((pc[-3]) & 0xff000000)}.
6972 Generate code for the 16-bit Thumb instruction set. The default is to
6973 use the 32-bit ARM instruction set.
6976 @opindex mtpcs-frame
6977 Generate a stack frame that is compliant with the Thumb Procedure Call
6978 Standard for all non-leaf functions. (A leaf function is one that does
6979 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6981 @item -mtpcs-leaf-frame
6982 @opindex mtpcs-leaf-frame
6983 Generate a stack frame that is compliant with the Thumb Procedure Call
6984 Standard for all leaf functions. (A leaf function is one that does
6985 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6987 @item -mcallee-super-interworking
6988 @opindex mcallee-super-interworking
6989 Gives all externally visible functions in the file being compiled an ARM
6990 instruction set header which switches to Thumb mode before executing the
6991 rest of the function. This allows these functions to be called from
6992 non-interworking code.
6994 @item -mcaller-super-interworking
6995 @opindex mcaller-super-interworking
6996 Allows calls via function pointers (including virtual functions) to
6997 execute correctly regardless of whether the target code has been
6998 compiled for interworking or not. There is a small overhead in the cost
6999 of executing a function pointer if this option is enabled.
7004 @subsection AVR Options
7007 These options are defined for AVR implementations:
7010 @item -mmcu=@var{mcu}
7012 Specify ATMEL AVR instruction set or MCU type.
7014 Instruction set avr1 is for the minimal AVR core, not supported by the C
7015 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
7016 attiny11, attiny12, attiny15, attiny28).
7018 Instruction set avr2 (default) is for the classic AVR core with up to
7019 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
7020 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
7021 at90c8534, at90s8535).
7023 Instruction set avr3 is for the classic AVR core with up to 128K program
7024 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
7026 Instruction set avr4 is for the enhanced AVR core with up to 8K program
7027 memory space (MCU types: atmega8, atmega83, atmega85).
7029 Instruction set avr5 is for the enhanced AVR core with up to 128K program
7030 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
7031 atmega64, atmega128, at43usb355, at94k).
7035 Output instruction sizes to the asm file.
7037 @item -minit-stack=@var{N}
7038 @opindex minit-stack
7039 Specify the initial stack address, which may be a symbol or numeric value,
7040 @samp{__stack} is the default.
7042 @item -mno-interrupts
7043 @opindex mno-interrupts
7044 Generated code is not compatible with hardware interrupts.
7045 Code size will be smaller.
7047 @item -mcall-prologues
7048 @opindex mcall-prologues
7049 Functions prologues/epilogues expanded as call to appropriate
7050 subroutines. Code size will be smaller.
7052 @item -mno-tablejump
7053 @opindex mno-tablejump
7054 Do not generate tablejump insns which sometimes increase code size.
7057 @opindex mtiny-stack
7058 Change only the low 8 bits of the stack pointer.
7062 Assume int to be 8 bit integer. This affects the sizes of all types: A
7063 char will be 1 byte, an int will be 1 byte, an long will be 2 bytes
7064 and long long will be 4 bytes. Please note that this option does not
7065 comply to the C standards, but it will provide you with smaller code
7070 @subsection CRIS Options
7071 @cindex CRIS Options
7073 These options are defined specifically for the CRIS ports.
7076 @item -march=@var{architecture-type}
7077 @itemx -mcpu=@var{architecture-type}
7080 Generate code for the specified architecture. The choices for
7081 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
7082 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
7083 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
7086 @item -mtune=@var{architecture-type}
7088 Tune to @var{architecture-type} everything applicable about the generated
7089 code, except for the ABI and the set of available instructions. The
7090 choices for @var{architecture-type} are the same as for
7091 @option{-march=@var{architecture-type}}.
7093 @item -mmax-stack-frame=@var{n}
7094 @opindex mmax-stack-frame
7095 Warn when the stack frame of a function exceeds @var{n} bytes.
7097 @item -melinux-stacksize=@var{n}
7098 @opindex melinux-stacksize
7099 Only available with the @samp{cris-axis-aout} target. Arranges for
7100 indications in the program to the kernel loader that the stack of the
7101 program should be set to @var{n} bytes.
7107 The options @option{-metrax4} and @option{-metrax100} are synonyms for
7108 @option{-march=v3} and @option{-march=v8} respectively.
7110 @item -mmul-bug-workaround
7111 @itemx -mno-mul-bug-workaround
7112 @opindex mmul-bug-workaround
7113 @opindex mno-mul-bug-workaround
7114 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
7115 models where it applies. This option is active by default.
7119 Enable CRIS-specific verbose debug-related information in the assembly
7120 code. This option also has the effect to turn off the @samp{#NO_APP}
7121 formatted-code indicator to the assembler at the beginning of the
7126 Do not use condition-code results from previous instruction; always emit
7127 compare and test instructions before use of condition codes.
7129 @item -mno-side-effects
7130 @opindex mno-side-effects
7131 Do not emit instructions with side-effects in addressing modes other than
7135 @itemx -mno-stack-align
7137 @itemx -mno-data-align
7138 @itemx -mconst-align
7139 @itemx -mno-const-align
7140 @opindex mstack-align
7141 @opindex mno-stack-align
7142 @opindex mdata-align
7143 @opindex mno-data-align
7144 @opindex mconst-align
7145 @opindex mno-const-align
7146 These options (no-options) arranges (eliminate arrangements) for the
7147 stack-frame, individual data and constants to be aligned for the maximum
7148 single data access size for the chosen CPU model. The default is to
7149 arrange for 32-bit alignment. ABI details such as structure layout are
7150 not affected by these options.
7158 Similar to the stack- data- and const-align options above, these options
7159 arrange for stack-frame, writable data and constants to all be 32-bit,
7160 16-bit or 8-bit aligned. The default is 32-bit alignment.
7162 @item -mno-prologue-epilogue
7163 @itemx -mprologue-epilogue
7164 @opindex mno-prologue-epilogue
7165 @opindex mprologue-epilogue
7166 With @option{-mno-prologue-epilogue}, the normal function prologue and
7167 epilogue that sets up the stack-frame are omitted and no return
7168 instructions or return sequences are generated in the code. Use this
7169 option only together with visual inspection of the compiled code: no
7170 warnings or errors are generated when call-saved registers must be saved,
7171 or storage for local variable needs to be allocated.
7177 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
7178 instruction sequences that load addresses for functions from the PLT part
7179 of the GOT rather than (traditional on other architectures) calls to the
7180 PLT@. The default is @option{-mgotplt}.
7184 Legacy no-op option only recognized with the cris-axis-aout target.
7188 Legacy no-op option only recognized with the cris-axis-elf and
7189 cris-axis-linux-gnu targets.
7193 Only recognized with the cris-axis-aout target, where it selects a
7194 GNU/linux-like multilib, include files and instruction set for
7199 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
7203 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
7204 to link with input-output functions from a simulator library. Code,
7205 initialized data and zero-initialized data are allocated consecutively.
7209 Like @option{-sim}, but pass linker options to locate initialized data at
7210 0x40000000 and zero-initialized data at 0x80000000.
7213 @node Darwin Options
7214 @subsection Darwin Options
7215 @cindex Darwin options
7217 These options are defined for all architectures running the Darwin operating
7220 FSF GCC on Darwin does not create ``fat'' object files; it will create
7221 an object file for the single architecture that it was built to
7222 target. Apple's GCC on Darwin does create ``fat'' files if multiple
7223 @option{-arch} options are used; it does so by running the compiler or
7224 linker multiple times and joining the results together with
7227 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
7228 @samp{i686}) is determined by the flags that specify the ISA
7229 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
7230 @option{-force_cpusubtype_ALL} option can be used to override this.
7232 The Darwin tools vary in their behaviour when presented with an ISA
7233 mismatch. The assembler, @file{as}, will only permit instructions to
7234 be used that are valid for the subtype of the file it is generating,
7235 so you cannot put 64-bit instructions in an @samp{ppc750} object file.
7236 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
7237 and print an error if asked to create a shared library with a less
7238 restrictive subtype than its input files (for instance, trying to put
7239 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
7240 for executables, @file{ld}, will quietly give the executable the most
7241 restrictive subtype of any of its input files.
7246 Add the framework directory @var{dir} to the head of the list of
7247 directories to be searched for header files. These directories are
7248 interleaved with those specified by @option{-I} options and are
7249 scanned in a left-to-right order.
7251 A framework directory is a directory with frameworks in it. A
7252 framework is a directory with a @samp{"Headers"} and/or
7253 @samp{"PrivateHeaders"} directory contained directly in it that ends
7254 in @samp{".framework"}. The name of a framework is the name of this
7255 directory excluding the @samp{".framework"}. Headers associated with
7256 the framework are found in one of those two directories, with
7257 @samp{"Headers"} being searched first. A subframework is a framework
7258 directory that is in a framework's @samp{"Frameworks"} directory.
7259 Includes of subframework headers can only appear in a header of a
7260 framework that contains the subframework, or in a sibling subframework
7261 header. Two subframeworks are siblings if they occur in the same
7262 framework. A subframework should not have the same name as a
7263 framework, a warning will be issued if this is violated. Currently a
7264 subframework cannot have subframeworks, in the future, the mechanism
7265 may be extended to support this. The standard frameworks can be found
7266 in @samp{"/System/Library/Frameworks"} and
7267 @samp{"/Library/Frameworks"}. An example include looks like
7268 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
7269 the name of the framework and header.h is found in the
7270 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
7274 Emit debugging information for symbols that are used. For STABS
7275 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
7276 This is by default ON@.
7280 Emit debugging information for all symbols and types.
7282 @item -mone-byte-bool
7283 @opindex -mone-byte-bool
7284 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
7285 By default @samp{sizeof(bool)} is @samp{4} when compiling for
7286 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
7287 option has no effect on x86.
7289 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
7290 to generate code that is not binary compatible with code generated
7291 without that switch. Using this switch may require recompiling all
7292 other modules in a program, including system libraries. Use this
7293 switch to conform to a non-default data model.
7295 @item -mfix-and-continue
7296 @itemx -ffix-and-continue
7297 @itemx -findirect-data
7298 @opindex mfix-and-continue
7299 @opindex ffix-and-continue
7300 @opindex findirect-data
7301 Generate code suitable for fast turn around development. Needed to
7302 enable gdb to dynamically load @code{.o} files into already running
7303 programs. @option{-findirect-data} and @option{-ffix-and-continue}
7304 are provided for backwards compatibility.
7308 Loads all members of static archive libraries.
7309 See man ld(1) for more information.
7311 @item -arch_errors_fatal
7312 @opindex arch_errors_fatal
7313 Cause the errors having to do with files that have the wrong architecture
7317 @opindex bind_at_load
7318 Causes the output file to be marked such that the dynamic linker will
7319 bind all undefined references when the file is loaded or launched.
7323 Produce a Mach-o bundle format file.
7324 See man ld(1) for more information.
7326 @item -bundle_loader @var{executable}
7327 @opindex bundle_loader
7328 This option specifies the @var{executable} that will be loading the build
7329 output file being linked. See man ld(1) for more information.
7332 @opindex -dynamiclib
7333 When passed this option, GCC will produce a dynamic library instead of
7334 an executable when linking, using the Darwin @file{libtool} command.
7336 @item -force_cpusubtype_ALL
7337 @opindex -force_cpusubtype_ALL
7338 This causes GCC's output file to have the @var{ALL} subtype, instead of
7339 one controlled by the @option{-mcpu} or @option{-march} option.
7341 @item -allowable_client @var{client_name}
7343 @itemx -compatibility_version
7344 @itemx -current_version
7346 @itemx -dependency-file
7348 @itemx -dylinker_install_name
7350 @itemx -exported_symbols_list
7352 @itemx -flat_namespace
7353 @itemx -force_flat_namespace
7354 @itemx -headerpad_max_install_names
7357 @itemx -install_name
7358 @itemx -keep_private_externs
7359 @itemx -multi_module
7360 @itemx -multiply_defined
7361 @itemx -multiply_defined_unused
7363 @itemx -no_dead_strip_inits_and_terms
7364 @itemx -nofixprebinding
7367 @itemx -noseglinkedit
7368 @itemx -pagezero_size
7370 @itemx -prebind_all_twolevel_modules
7371 @itemx -private_bundle
7372 @itemx -read_only_relocs
7374 @itemx -sectobjectsymbols
7378 @itemx -sectobjectsymbols
7381 @itemx -segs_read_only_addr
7382 @itemx -segs_read_write_addr
7383 @itemx -seg_addr_table
7384 @itemx -seg_addr_table_filename
7387 @itemx -segs_read_only_addr
7388 @itemx -segs_read_write_addr
7389 @itemx -single_module
7392 @itemx -sub_umbrella
7393 @itemx -twolevel_namespace
7396 @itemx -unexported_symbols_list
7397 @itemx -weak_reference_mismatches
7400 @opindex allowable_client
7401 @opindex client_name
7402 @opindex compatibility_version
7403 @opindex current_version
7405 @opindex dependency-file
7407 @opindex dylinker_install_name
7409 @opindex exported_symbols_list
7411 @opindex flat_namespace
7412 @opindex force_flat_namespace
7413 @opindex headerpad_max_install_names
7416 @opindex install_name
7417 @opindex keep_private_externs
7418 @opindex multi_module
7419 @opindex multiply_defined
7420 @opindex multiply_defined_unused
7422 @opindex no_dead_strip_inits_and_terms
7423 @opindex nofixprebinding
7424 @opindex nomultidefs
7426 @opindex noseglinkedit
7427 @opindex pagezero_size
7429 @opindex prebind_all_twolevel_modules
7430 @opindex private_bundle
7431 @opindex read_only_relocs
7433 @opindex sectobjectsymbols
7437 @opindex sectobjectsymbols
7440 @opindex segs_read_only_addr
7441 @opindex segs_read_write_addr
7442 @opindex seg_addr_table
7443 @opindex seg_addr_table_filename
7444 @opindex seglinkedit
7446 @opindex segs_read_only_addr
7447 @opindex segs_read_write_addr
7448 @opindex single_module
7450 @opindex sub_library
7451 @opindex sub_umbrella
7452 @opindex twolevel_namespace
7455 @opindex unexported_symbols_list
7456 @opindex weak_reference_mismatches
7457 @opindex whatsloaded
7459 These options are passed to the Darwin linker. The Darwin linker man page
7460 describes them in detail.
7463 @node DEC Alpha Options
7464 @subsection DEC Alpha Options
7466 These @samp{-m} options are defined for the DEC Alpha implementations:
7469 @item -mno-soft-float
7471 @opindex mno-soft-float
7472 @opindex msoft-float
7473 Use (do not use) the hardware floating-point instructions for
7474 floating-point operations. When @option{-msoft-float} is specified,
7475 functions in @file{libgcc.a} will be used to perform floating-point
7476 operations. Unless they are replaced by routines that emulate the
7477 floating-point operations, or compiled in such a way as to call such
7478 emulations routines, these routines will issue floating-point
7479 operations. If you are compiling for an Alpha without floating-point
7480 operations, you must ensure that the library is built so as not to call
7483 Note that Alpha implementations without floating-point operations are
7484 required to have floating-point registers.
7489 @opindex mno-fp-regs
7490 Generate code that uses (does not use) the floating-point register set.
7491 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
7492 register set is not used, floating point operands are passed in integer
7493 registers as if they were integers and floating-point results are passed
7494 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
7495 so any function with a floating-point argument or return value called by code
7496 compiled with @option{-mno-fp-regs} must also be compiled with that
7499 A typical use of this option is building a kernel that does not use,
7500 and hence need not save and restore, any floating-point registers.
7504 The Alpha architecture implements floating-point hardware optimized for
7505 maximum performance. It is mostly compliant with the IEEE floating
7506 point standard. However, for full compliance, software assistance is
7507 required. This option generates code fully IEEE compliant code
7508 @emph{except} that the @var{inexact-flag} is not maintained (see below).
7509 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
7510 defined during compilation. The resulting code is less efficient but is
7511 able to correctly support denormalized numbers and exceptional IEEE
7512 values such as not-a-number and plus/minus infinity. Other Alpha
7513 compilers call this option @option{-ieee_with_no_inexact}.
7515 @item -mieee-with-inexact
7516 @opindex mieee-with-inexact
7517 This is like @option{-mieee} except the generated code also maintains
7518 the IEEE @var{inexact-flag}. Turning on this option causes the
7519 generated code to implement fully-compliant IEEE math. In addition to
7520 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
7521 macro. On some Alpha implementations the resulting code may execute
7522 significantly slower than the code generated by default. Since there is
7523 very little code that depends on the @var{inexact-flag}, you should
7524 normally not specify this option. Other Alpha compilers call this
7525 option @option{-ieee_with_inexact}.
7527 @item -mfp-trap-mode=@var{trap-mode}
7528 @opindex mfp-trap-mode
7529 This option controls what floating-point related traps are enabled.
7530 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
7531 The trap mode can be set to one of four values:
7535 This is the default (normal) setting. The only traps that are enabled
7536 are the ones that cannot be disabled in software (e.g., division by zero
7540 In addition to the traps enabled by @samp{n}, underflow traps are enabled
7544 Like @samp{su}, but the instructions are marked to be safe for software
7545 completion (see Alpha architecture manual for details).
7548 Like @samp{su}, but inexact traps are enabled as well.
7551 @item -mfp-rounding-mode=@var{rounding-mode}
7552 @opindex mfp-rounding-mode
7553 Selects the IEEE rounding mode. Other Alpha compilers call this option
7554 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
7559 Normal IEEE rounding mode. Floating point numbers are rounded towards
7560 the nearest machine number or towards the even machine number in case
7564 Round towards minus infinity.
7567 Chopped rounding mode. Floating point numbers are rounded towards zero.
7570 Dynamic rounding mode. A field in the floating point control register
7571 (@var{fpcr}, see Alpha architecture reference manual) controls the
7572 rounding mode in effect. The C library initializes this register for
7573 rounding towards plus infinity. Thus, unless your program modifies the
7574 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
7577 @item -mtrap-precision=@var{trap-precision}
7578 @opindex mtrap-precision
7579 In the Alpha architecture, floating point traps are imprecise. This
7580 means without software assistance it is impossible to recover from a
7581 floating trap and program execution normally needs to be terminated.
7582 GCC can generate code that can assist operating system trap handlers
7583 in determining the exact location that caused a floating point trap.
7584 Depending on the requirements of an application, different levels of
7585 precisions can be selected:
7589 Program precision. This option is the default and means a trap handler
7590 can only identify which program caused a floating point exception.
7593 Function precision. The trap handler can determine the function that
7594 caused a floating point exception.
7597 Instruction precision. The trap handler can determine the exact
7598 instruction that caused a floating point exception.
7601 Other Alpha compilers provide the equivalent options called
7602 @option{-scope_safe} and @option{-resumption_safe}.
7604 @item -mieee-conformant
7605 @opindex mieee-conformant
7606 This option marks the generated code as IEEE conformant. You must not
7607 use this option unless you also specify @option{-mtrap-precision=i} and either
7608 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
7609 is to emit the line @samp{.eflag 48} in the function prologue of the
7610 generated assembly file. Under DEC Unix, this has the effect that
7611 IEEE-conformant math library routines will be linked in.
7613 @item -mbuild-constants
7614 @opindex mbuild-constants
7615 Normally GCC examines a 32- or 64-bit integer constant to
7616 see if it can construct it from smaller constants in two or three
7617 instructions. If it cannot, it will output the constant as a literal and
7618 generate code to load it from the data segment at runtime.
7620 Use this option to require GCC to construct @emph{all} integer constants
7621 using code, even if it takes more instructions (the maximum is six).
7623 You would typically use this option to build a shared library dynamic
7624 loader. Itself a shared library, it must relocate itself in memory
7625 before it can find the variables and constants in its own data segment.
7631 Select whether to generate code to be assembled by the vendor-supplied
7632 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
7650 Indicate whether GCC should generate code to use the optional BWX,
7651 CIX, FIX and MAX instruction sets. The default is to use the instruction
7652 sets supported by the CPU type specified via @option{-mcpu=} option or that
7653 of the CPU on which GCC was built if none was specified.
7658 @opindex mfloat-ieee
7659 Generate code that uses (does not use) VAX F and G floating point
7660 arithmetic instead of IEEE single and double precision.
7662 @item -mexplicit-relocs
7663 @itemx -mno-explicit-relocs
7664 @opindex mexplicit-relocs
7665 @opindex mno-explicit-relocs
7666 Older Alpha assemblers provided no way to generate symbol relocations
7667 except via assembler macros. Use of these macros does not allow
7668 optimal instruction scheduling. GNU binutils as of version 2.12
7669 supports a new syntax that allows the compiler to explicitly mark
7670 which relocations should apply to which instructions. This option
7671 is mostly useful for debugging, as GCC detects the capabilities of
7672 the assembler when it is built and sets the default accordingly.
7676 @opindex msmall-data
7677 @opindex mlarge-data
7678 When @option{-mexplicit-relocs} is in effect, static data is
7679 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
7680 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
7681 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
7682 16-bit relocations off of the @code{$gp} register. This limits the
7683 size of the small data area to 64KB, but allows the variables to be
7684 directly accessed via a single instruction.
7686 The default is @option{-mlarge-data}. With this option the data area
7687 is limited to just below 2GB@. Programs that require more than 2GB of
7688 data must use @code{malloc} or @code{mmap} to allocate the data in the
7689 heap instead of in the program's data segment.
7691 When generating code for shared libraries, @option{-fpic} implies
7692 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
7696 @opindex msmall-text
7697 @opindex mlarge-text
7698 When @option{-msmall-text} is used, the compiler assumes that the
7699 code of the entire program (or shared library) fits in 4MB, and is
7700 thus reachable with a branch instruction. When @option{-msmall-data}
7701 is used, the compiler can assume that all local symbols share the
7702 same @code{$gp} value, and thus reduce the number of instructions
7703 required for a function call from 4 to 1.
7705 The default is @option{-mlarge-text}.
7707 @item -mcpu=@var{cpu_type}
7709 Set the instruction set and instruction scheduling parameters for
7710 machine type @var{cpu_type}. You can specify either the @samp{EV}
7711 style name or the corresponding chip number. GCC supports scheduling
7712 parameters for the EV4, EV5 and EV6 family of processors and will
7713 choose the default values for the instruction set from the processor
7714 you specify. If you do not specify a processor type, GCC will default
7715 to the processor on which the compiler was built.
7717 Supported values for @var{cpu_type} are
7723 Schedules as an EV4 and has no instruction set extensions.
7727 Schedules as an EV5 and has no instruction set extensions.
7731 Schedules as an EV5 and supports the BWX extension.
7736 Schedules as an EV5 and supports the BWX and MAX extensions.
7740 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
7744 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
7747 @item -mtune=@var{cpu_type}
7749 Set only the instruction scheduling parameters for machine type
7750 @var{cpu_type}. The instruction set is not changed.
7752 @item -mmemory-latency=@var{time}
7753 @opindex mmemory-latency
7754 Sets the latency the scheduler should assume for typical memory
7755 references as seen by the application. This number is highly
7756 dependent on the memory access patterns used by the application
7757 and the size of the external cache on the machine.
7759 Valid options for @var{time} are
7763 A decimal number representing clock cycles.
7769 The compiler contains estimates of the number of clock cycles for
7770 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
7771 (also called Dcache, Scache, and Bcache), as well as to main memory.
7772 Note that L3 is only valid for EV5.
7777 @node DEC Alpha/VMS Options
7778 @subsection DEC Alpha/VMS Options
7780 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
7783 @item -mvms-return-codes
7784 @opindex mvms-return-codes
7785 Return VMS condition codes from main. The default is to return POSIX
7786 style condition (e.g.@ error) codes.
7790 @subsection FRV Options
7797 Only use the first 32 general purpose registers.
7802 Use all 64 general purpose registers.
7807 Use only the first 32 floating point registers.
7812 Use all 64 floating point registers
7815 @opindex mhard-float
7817 Use hardware instructions for floating point operations.
7820 @opindex msoft-float
7822 Use library routines for floating point operations.
7827 Dynamically allocate condition code registers.
7832 Do not try to dynamically allocate condition code registers, only
7833 use @code{icc0} and @code{fcc0}.
7838 Change ABI to use double word insns.
7843 Do not use double word instructions.
7848 Use floating point double instructions.
7853 Do not use floating point double instructions.
7858 Use media instructions.
7863 Do not use media instructions.
7868 Use multiply and add/subtract instructions.
7873 Do not use multiply and add/subtract instructions.
7878 Select the FDPIC ABI, that uses function descriptors to represent
7879 pointers to functions. Without any PIC/PIE-related options, it
7880 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
7881 assumes GOT entries and small data are within a 12-bit range from the
7882 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
7883 are computed with 32 bits.
7886 @opindex minline-plt
7888 Enable inlining of PLT entries in function calls to functions that are
7889 not known to bind locally. It has no effect without @option{-mfdpic}.
7890 It's enabled by default if optimizing for speed and compiling for
7891 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
7892 optimization option such as @option{-O3} or above is present in the
7898 Assume a large TLS segment when generating thread-local code.
7903 Do not assume a large TLS segment when generating thread-local code.
7908 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
7909 that is known to be in read-only sections. It's enabled by default,
7910 except for @option{-fpic} or @option{-fpie}: even though it may help
7911 make the global offset table smaller, it trades 1 instruction for 4.
7912 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
7913 one of which may be shared by multiple symbols, and it avoids the need
7914 for a GOT entry for the referenced symbol, so it's more likely to be a
7915 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
7917 @item -multilib-library-pic
7918 @opindex multilib-library-pic
7920 Link with the (library, not FD) pic libraries. It's implied by
7921 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
7922 @option{-fpic} without @option{-mfdpic}. You should never have to use
7928 Follow the EABI requirement of always creating a frame pointer whenever
7929 a stack frame is allocated. This option is enabled by default and can
7930 be disabled with @option{-mno-linked-fp}.
7933 @opindex mlong-calls
7935 Use indirect addressing to call functions outside the current
7936 compilation unit. This allows the functions to be placed anywhere
7937 within the 32-bit address space.
7939 @item -malign-labels
7940 @opindex malign-labels
7942 Try to align labels to an 8-byte boundary by inserting nops into the
7943 previous packet. This option only has an effect when VLIW packing
7944 is enabled. It doesn't create new packets; it merely adds nops to
7948 @opindex mlibrary-pic
7950 Generate position-independent EABI code.
7955 Use only the first four media accumulator registers.
7960 Use all eight media accumulator registers.
7965 Pack VLIW instructions.
7970 Do not pack VLIW instructions.
7975 Do not mark ABI switches in e_flags.
7980 Enable the use of conditional-move instructions (default).
7982 This switch is mainly for debugging the compiler and will likely be removed
7983 in a future version.
7985 @item -mno-cond-move
7986 @opindex mno-cond-move
7988 Disable the use of conditional-move instructions.
7990 This switch is mainly for debugging the compiler and will likely be removed
7991 in a future version.
7996 Enable the use of conditional set instructions (default).
7998 This switch is mainly for debugging the compiler and will likely be removed
7999 in a future version.
8004 Disable the use of conditional set instructions.
8006 This switch is mainly for debugging the compiler and will likely be removed
8007 in a future version.
8012 Enable the use of conditional execution (default).
8014 This switch is mainly for debugging the compiler and will likely be removed
8015 in a future version.
8017 @item -mno-cond-exec
8018 @opindex mno-cond-exec
8020 Disable the use of conditional execution.
8022 This switch is mainly for debugging the compiler and will likely be removed
8023 in a future version.
8026 @opindex mvliw-branch
8028 Run a pass to pack branches into VLIW instructions (default).
8030 This switch is mainly for debugging the compiler and will likely be removed
8031 in a future version.
8033 @item -mno-vliw-branch
8034 @opindex mno-vliw-branch
8036 Do not run a pass to pack branches into VLIW instructions.
8038 This switch is mainly for debugging the compiler and will likely be removed
8039 in a future version.
8041 @item -mmulti-cond-exec
8042 @opindex mmulti-cond-exec
8044 Enable optimization of @code{&&} and @code{||} in conditional execution
8047 This switch is mainly for debugging the compiler and will likely be removed
8048 in a future version.
8050 @item -mno-multi-cond-exec
8051 @opindex mno-multi-cond-exec
8053 Disable optimization of @code{&&} and @code{||} in conditional execution.
8055 This switch is mainly for debugging the compiler and will likely be removed
8056 in a future version.
8058 @item -mnested-cond-exec
8059 @opindex mnested-cond-exec
8061 Enable nested conditional execution optimizations (default).
8063 This switch is mainly for debugging the compiler and will likely be removed
8064 in a future version.
8066 @item -mno-nested-cond-exec
8067 @opindex mno-nested-cond-exec
8069 Disable nested conditional execution optimizations.
8071 This switch is mainly for debugging the compiler and will likely be removed
8072 in a future version.
8074 @item -mtomcat-stats
8075 @opindex mtomcat-stats
8077 Cause gas to print out tomcat statistics.
8079 @item -mcpu=@var{cpu}
8082 Select the processor type for which to generate code. Possible values are
8083 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
8084 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
8088 @node H8/300 Options
8089 @subsection H8/300 Options
8091 These @samp{-m} options are defined for the H8/300 implementations:
8096 Shorten some address references at link time, when possible; uses the
8097 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8098 ld, Using ld}, for a fuller description.
8102 Generate code for the H8/300H@.
8106 Generate code for the H8S@.
8110 Generate code for the H8S and H8/300H in the normal mode. This switch
8111 must be used either with @option{-mh} or @option{-ms}.
8115 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
8119 Make @code{int} data 32 bits by default.
8123 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
8124 The default for the H8/300H and H8S is to align longs and floats on 4
8126 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8127 This option has no effect on the H8/300.
8131 @subsection HPPA Options
8132 @cindex HPPA Options
8134 These @samp{-m} options are defined for the HPPA family of computers:
8137 @item -march=@var{architecture-type}
8139 Generate code for the specified architecture. The choices for
8140 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
8141 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
8142 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
8143 architecture option for your machine. Code compiled for lower numbered
8144 architectures will run on higher numbered architectures, but not the
8148 @itemx -mpa-risc-1-1
8149 @itemx -mpa-risc-2-0
8150 @opindex mpa-risc-1-0
8151 @opindex mpa-risc-1-1
8152 @opindex mpa-risc-2-0
8153 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
8156 @opindex mbig-switch
8157 Generate code suitable for big switch tables. Use this option only if
8158 the assembler/linker complain about out of range branches within a switch
8161 @item -mjump-in-delay
8162 @opindex mjump-in-delay
8163 Fill delay slots of function calls with unconditional jump instructions
8164 by modifying the return pointer for the function call to be the target
8165 of the conditional jump.
8167 @item -mdisable-fpregs
8168 @opindex mdisable-fpregs
8169 Prevent floating point registers from being used in any manner. This is
8170 necessary for compiling kernels which perform lazy context switching of
8171 floating point registers. If you use this option and attempt to perform
8172 floating point operations, the compiler will abort.
8174 @item -mdisable-indexing
8175 @opindex mdisable-indexing
8176 Prevent the compiler from using indexing address modes. This avoids some
8177 rather obscure problems when compiling MIG generated code under MACH@.
8179 @item -mno-space-regs
8180 @opindex mno-space-regs
8181 Generate code that assumes the target has no space registers. This allows
8182 GCC to generate faster indirect calls and use unscaled index address modes.
8184 Such code is suitable for level 0 PA systems and kernels.
8186 @item -mfast-indirect-calls
8187 @opindex mfast-indirect-calls
8188 Generate code that assumes calls never cross space boundaries. This
8189 allows GCC to emit code which performs faster indirect calls.
8191 This option will not work in the presence of shared libraries or nested
8194 @item -mfixed-range=@var{register-range}
8195 @opindex mfixed-range
8196 Generate code treating the given register range as fixed registers.
8197 A fixed register is one that the register allocator can not use. This is
8198 useful when compiling kernel code. A register range is specified as
8199 two registers separated by a dash. Multiple register ranges can be
8200 specified separated by a comma.
8202 @item -mlong-load-store
8203 @opindex mlong-load-store
8204 Generate 3-instruction load and store sequences as sometimes required by
8205 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
8208 @item -mportable-runtime
8209 @opindex mportable-runtime
8210 Use the portable calling conventions proposed by HP for ELF systems.
8214 Enable the use of assembler directives only GAS understands.
8216 @item -mschedule=@var{cpu-type}
8218 Schedule code according to the constraints for the machine type
8219 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
8220 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
8221 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
8222 proper scheduling option for your machine. The default scheduling is
8226 @opindex mlinker-opt
8227 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
8228 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
8229 linkers in which they give bogus error messages when linking some programs.
8232 @opindex msoft-float
8233 Generate output containing library calls for floating point.
8234 @strong{Warning:} the requisite libraries are not available for all HPPA
8235 targets. Normally the facilities of the machine's usual C compiler are
8236 used, but this cannot be done directly in cross-compilation. You must make
8237 your own arrangements to provide suitable library functions for
8238 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
8239 does provide software floating point support.
8241 @option{-msoft-float} changes the calling convention in the output file;
8242 therefore, it is only useful if you compile @emph{all} of a program with
8243 this option. In particular, you need to compile @file{libgcc.a}, the
8244 library that comes with GCC, with @option{-msoft-float} in order for
8249 Generate the predefine, @code{_SIO}, for server IO@. The default is
8250 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
8251 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
8252 options are available under HP-UX and HI-UX@.
8256 Use GNU ld specific options. This passes @option{-shared} to ld when
8257 building a shared library. It is the default when GCC is configured,
8258 explicitly or implicitly, with the GNU linker. This option does not
8259 have any affect on which ld is called, it only changes what parameters
8260 are passed to that ld. The ld that is called is determined by the
8261 @option{--with-ld} configure option, GCC's program search path, and
8262 finally by the user's @env{PATH}. The linker used by GCC can be printed
8263 using @samp{which `gcc -print-prog-name=ld`}.
8267 Use HP ld specific options. This passes @option{-b} to ld when building
8268 a shared library and passes @option{+Accept TypeMismatch} to ld on all
8269 links. It is the default when GCC is configured, explicitly or
8270 implicitly, with the HP linker. This option does not have any affect on
8271 which ld is called, it only changes what parameters are passed to that
8272 ld. The ld that is called is determined by the @option{--with-ld}
8273 configure option, GCC's program search path, and finally by the user's
8274 @env{PATH}. The linker used by GCC can be printed using @samp{which
8275 `gcc -print-prog-name=ld`}.
8278 @opindex mno-long-calls
8279 Generate code that uses long call sequences. This ensures that a call
8280 is always able to reach linker generated stubs. The default is to generate
8281 long calls only when the distance from the call site to the beginning
8282 of the function or translation unit, as the case may be, exceeds a
8283 predefined limit set by the branch type being used. The limits for
8284 normal calls are 7,600,000 and 240,000 bytes, respectively for the
8285 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
8288 Distances are measured from the beginning of functions when using the
8289 @option{-ffunction-sections} option, or when using the @option{-mgas}
8290 and @option{-mno-portable-runtime} options together under HP-UX with
8293 It is normally not desirable to use this option as it will degrade
8294 performance. However, it may be useful in large applications,
8295 particularly when partial linking is used to build the application.
8297 The types of long calls used depends on the capabilities of the
8298 assembler and linker, and the type of code being generated. The
8299 impact on systems that support long absolute calls, and long pic
8300 symbol-difference or pc-relative calls should be relatively small.
8301 However, an indirect call is used on 32-bit ELF systems in pic code
8302 and it is quite long.
8304 @item -munix=@var{unix-std}
8306 Generate compiler predefines and select a startfile for the specified
8307 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
8308 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
8309 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
8310 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
8311 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
8314 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
8315 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
8316 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
8317 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
8318 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
8319 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
8321 It is @emph{important} to note that this option changes the interfaces
8322 for various library routines. It also affects the operational behavior
8323 of the C library. Thus, @emph{extreme} care is needed in using this
8326 Library code that is intended to operate with more than one UNIX
8327 standard must test, set and restore the variable @var{__xpg4_extended_mask}
8328 as appropriate. Most GNU software doesn't provide this capability.
8332 Suppress the generation of link options to search libdld.sl when the
8333 @option{-static} option is specified on HP-UX 10 and later.
8337 The HP-UX implementation of setlocale in libc has a dependency on
8338 libdld.sl. There isn't an archive version of libdld.sl. Thus,
8339 when the @option{-static} option is specified, special link options
8340 are needed to resolve this dependency.
8342 On HP-UX 10 and later, the GCC driver adds the necessary options to
8343 link with libdld.sl when the @option{-static} option is specified.
8344 This causes the resulting binary to be dynamic. On the 64-bit port,
8345 the linkers generate dynamic binaries by default in any case. The
8346 @option{-nolibdld} option can be used to prevent the GCC driver from
8347 adding these link options.
8351 Add support for multithreading with the @dfn{dce thread} library
8352 under HP-UX@. This option sets flags for both the preprocessor and
8356 @node i386 and x86-64 Options
8357 @subsection Intel 386 and AMD x86-64 Options
8358 @cindex i386 Options
8359 @cindex x86-64 Options
8360 @cindex Intel 386 Options
8361 @cindex AMD x86-64 Options
8363 These @samp{-m} options are defined for the i386 and x86-64 family of
8367 @item -mtune=@var{cpu-type}
8369 Tune to @var{cpu-type} everything applicable about the generated code, except
8370 for the ABI and the set of available instructions. The choices for
8374 Original Intel's i386 CPU@.
8376 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
8378 Intel Pentium CPU with no MMX support.
8380 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
8381 @item i686, pentiumpro
8382 Intel PentiumPro CPU@.
8384 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
8385 @item pentium3, pentium3m
8386 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
8389 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
8390 support. Used by Centrino notebooks.
8391 @item pentium4, pentium4m
8392 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
8394 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
8397 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
8398 SSE2 and SSE3 instruction set support.
8400 AMD K6 CPU with MMX instruction set support.
8402 Improved versions of AMD K6 CPU with MMX and 3dNOW! instruction set support.
8403 @item athlon, athlon-tbird
8404 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and SSE prefetch instructions
8406 @item athlon-4, athlon-xp, athlon-mp
8407 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and full SSE
8408 instruction set support.
8409 @item k8, opteron, athlon64, athlon-fx
8410 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
8411 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW! and 64-bit instruction set extensions.)
8413 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
8416 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!
8417 instruction set support.
8419 Via C3 CPU with MMX and 3dNOW! instruction set support. (No scheduling is
8420 implemented for this chip.)
8422 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
8423 implemented for this chip.)
8426 While picking a specific @var{cpu-type} will schedule things appropriately
8427 for that particular chip, the compiler will not generate any code that
8428 does not run on the i386 without the @option{-march=@var{cpu-type}} option
8431 @item -march=@var{cpu-type}
8433 Generate instructions for the machine type @var{cpu-type}. The choices
8434 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
8435 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
8437 @item -mcpu=@var{cpu-type}
8439 A deprecated synonym for @option{-mtune}.
8448 @opindex mpentiumpro
8449 These options are synonyms for @option{-mtune=i386}, @option{-mtune=i486},
8450 @option{-mtune=pentium}, and @option{-mtune=pentiumpro} respectively.
8451 These synonyms are deprecated.
8453 @item -mfpmath=@var{unit}
8455 Generate floating point arithmetics for selected unit @var{unit}. The choices
8460 Use the standard 387 floating point coprocessor present majority of chips and
8461 emulated otherwise. Code compiled with this option will run almost everywhere.
8462 The temporary results are computed in 80bit precision instead of precision
8463 specified by the type resulting in slightly different results compared to most
8464 of other chips. See @option{-ffloat-store} for more detailed description.
8466 This is the default choice for i386 compiler.
8469 Use scalar floating point instructions present in the SSE instruction set.
8470 This instruction set is supported by Pentium3 and newer chips, in the AMD line
8471 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
8472 instruction set supports only single precision arithmetics, thus the double and
8473 extended precision arithmetics is still done using 387. Later version, present
8474 only in Pentium4 and the future AMD x86-64 chips supports double precision
8477 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
8478 or @option{-msse2} switches to enable SSE extensions and make this option
8479 effective. For the x86-64 compiler, these extensions are enabled by default.
8481 The resulting code should be considerably faster in the majority of cases and avoid
8482 the numerical instability problems of 387 code, but may break some existing
8483 code that expects temporaries to be 80bit.
8485 This is the default choice for the x86-64 compiler.
8488 Attempt to utilize both instruction sets at once. This effectively double the
8489 amount of available registers and on chips with separate execution units for
8490 387 and SSE the execution resources too. Use this option with care, as it is
8491 still experimental, because the GCC register allocator does not model separate
8492 functional units well resulting in instable performance.
8495 @item -masm=@var{dialect}
8496 @opindex masm=@var{dialect}
8497 Output asm instructions using selected @var{dialect}. Supported choices are
8498 @samp{intel} or @samp{att} (the default one).
8503 @opindex mno-ieee-fp
8504 Control whether or not the compiler uses IEEE floating point
8505 comparisons. These handle correctly the case where the result of a
8506 comparison is unordered.
8509 @opindex msoft-float
8510 Generate output containing library calls for floating point.
8511 @strong{Warning:} the requisite libraries are not part of GCC@.
8512 Normally the facilities of the machine's usual C compiler are used, but
8513 this can't be done directly in cross-compilation. You must make your
8514 own arrangements to provide suitable library functions for
8517 On machines where a function returns floating point results in the 80387
8518 register stack, some floating point opcodes may be emitted even if
8519 @option{-msoft-float} is used.
8521 @item -mno-fp-ret-in-387
8522 @opindex mno-fp-ret-in-387
8523 Do not use the FPU registers for return values of functions.
8525 The usual calling convention has functions return values of types
8526 @code{float} and @code{double} in an FPU register, even if there
8527 is no FPU@. The idea is that the operating system should emulate
8530 The option @option{-mno-fp-ret-in-387} causes such values to be returned
8531 in ordinary CPU registers instead.
8533 @item -mno-fancy-math-387
8534 @opindex mno-fancy-math-387
8535 Some 387 emulators do not support the @code{sin}, @code{cos} and
8536 @code{sqrt} instructions for the 387. Specify this option to avoid
8537 generating those instructions. This option is the default on FreeBSD,
8538 OpenBSD and NetBSD@. This option is overridden when @option{-march}
8539 indicates that the target cpu will always have an FPU and so the
8540 instruction will not need emulation. As of revision 2.6.1, these
8541 instructions are not generated unless you also use the
8542 @option{-funsafe-math-optimizations} switch.
8544 @item -malign-double
8545 @itemx -mno-align-double
8546 @opindex malign-double
8547 @opindex mno-align-double
8548 Control whether GCC aligns @code{double}, @code{long double}, and
8549 @code{long long} variables on a two word boundary or a one word
8550 boundary. Aligning @code{double} variables on a two word boundary will
8551 produce code that runs somewhat faster on a @samp{Pentium} at the
8552 expense of more memory.
8554 @strong{Warning:} if you use the @option{-malign-double} switch,
8555 structures containing the above types will be aligned differently than
8556 the published application binary interface specifications for the 386
8557 and will not be binary compatible with structures in code compiled
8558 without that switch.
8560 @item -m96bit-long-double
8561 @itemx -m128bit-long-double
8562 @opindex m96bit-long-double
8563 @opindex m128bit-long-double
8564 These switches control the size of @code{long double} type. The i386
8565 application binary interface specifies the size to be 96 bits,
8566 so @option{-m96bit-long-double} is the default in 32 bit mode.
8568 Modern architectures (Pentium and newer) would prefer @code{long double}
8569 to be aligned to an 8 or 16 byte boundary. In arrays or structures
8570 conforming to the ABI, this would not be possible. So specifying a
8571 @option{-m128bit-long-double} will align @code{long double}
8572 to a 16 byte boundary by padding the @code{long double} with an additional
8575 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
8576 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
8578 Notice that neither of these options enable any extra precision over the x87
8579 standard of 80 bits for a @code{long double}.
8581 @strong{Warning:} if you override the default value for your target ABI, the
8582 structures and arrays containing @code{long double} variables will change
8583 their size as well as function calling convention for function taking
8584 @code{long double} will be modified. Hence they will not be binary
8585 compatible with arrays or structures in code compiled without that switch.
8589 @itemx -mno-svr3-shlib
8590 @opindex msvr3-shlib
8591 @opindex mno-svr3-shlib
8592 Control whether GCC places uninitialized local variables into the
8593 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
8594 into @code{bss}. These options are meaningful only on System V Release 3.
8598 Use a different function-calling convention, in which functions that
8599 take a fixed number of arguments return with the @code{ret} @var{num}
8600 instruction, which pops their arguments while returning. This saves one
8601 instruction in the caller since there is no need to pop the arguments
8604 You can specify that an individual function is called with this calling
8605 sequence with the function attribute @samp{stdcall}. You can also
8606 override the @option{-mrtd} option by using the function attribute
8607 @samp{cdecl}. @xref{Function Attributes}.
8609 @strong{Warning:} this calling convention is incompatible with the one
8610 normally used on Unix, so you cannot use it if you need to call
8611 libraries compiled with the Unix compiler.
8613 Also, you must provide function prototypes for all functions that
8614 take variable numbers of arguments (including @code{printf});
8615 otherwise incorrect code will be generated for calls to those
8618 In addition, seriously incorrect code will result if you call a
8619 function with too many arguments. (Normally, extra arguments are
8620 harmlessly ignored.)
8622 @item -mregparm=@var{num}
8624 Control how many registers are used to pass integer arguments. By
8625 default, no registers are used to pass arguments, and at most 3
8626 registers can be used. You can control this behavior for a specific
8627 function by using the function attribute @samp{regparm}.
8628 @xref{Function Attributes}.
8630 @strong{Warning:} if you use this switch, and
8631 @var{num} is nonzero, then you must build all modules with the same
8632 value, including any libraries. This includes the system libraries and
8635 @item -mpreferred-stack-boundary=@var{num}
8636 @opindex mpreferred-stack-boundary
8637 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
8638 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
8639 the default is 4 (16 bytes or 128 bits), except when optimizing for code
8640 size (@option{-Os}), in which case the default is the minimum correct
8641 alignment (4 bytes for x86, and 8 bytes for x86-64).
8643 On Pentium and PentiumPro, @code{double} and @code{long double} values
8644 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
8645 suffer significant run time performance penalties. On Pentium III, the
8646 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
8647 penalties if it is not 16 byte aligned.
8649 To ensure proper alignment of this values on the stack, the stack boundary
8650 must be as aligned as that required by any value stored on the stack.
8651 Further, every function must be generated such that it keeps the stack
8652 aligned. Thus calling a function compiled with a higher preferred
8653 stack boundary from a function compiled with a lower preferred stack
8654 boundary will most likely misalign the stack. It is recommended that
8655 libraries that use callbacks always use the default setting.
8657 This extra alignment does consume extra stack space, and generally
8658 increases code size. Code that is sensitive to stack space usage, such
8659 as embedded systems and operating system kernels, may want to reduce the
8660 preferred alignment to @option{-mpreferred-stack-boundary=2}.
8678 These switches enable or disable the use of built-in functions that allow
8679 direct access to the MMX, SSE, SSE2, SSE3 and 3Dnow extensions of the
8682 @xref{X86 Built-in Functions}, for details of the functions enabled
8683 and disabled by these switches.
8685 To have SSE/SSE2 instructions generated automatically from floating-point
8686 code, see @option{-mfpmath=sse}.
8689 @itemx -mno-push-args
8691 @opindex mno-push-args
8692 Use PUSH operations to store outgoing parameters. This method is shorter
8693 and usually equally fast as method using SUB/MOV operations and is enabled
8694 by default. In some cases disabling it may improve performance because of
8695 improved scheduling and reduced dependencies.
8697 @item -maccumulate-outgoing-args
8698 @opindex maccumulate-outgoing-args
8699 If enabled, the maximum amount of space required for outgoing arguments will be
8700 computed in the function prologue. This is faster on most modern CPUs
8701 because of reduced dependencies, improved scheduling and reduced stack usage
8702 when preferred stack boundary is not equal to 2. The drawback is a notable
8703 increase in code size. This switch implies @option{-mno-push-args}.
8707 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
8708 on thread-safe exception handling must compile and link all code with the
8709 @option{-mthreads} option. When compiling, @option{-mthreads} defines
8710 @option{-D_MT}; when linking, it links in a special thread helper library
8711 @option{-lmingwthrd} which cleans up per thread exception handling data.
8713 @item -mno-align-stringops
8714 @opindex mno-align-stringops
8715 Do not align destination of inlined string operations. This switch reduces
8716 code size and improves performance in case the destination is already aligned,
8717 but GCC doesn't know about it.
8719 @item -minline-all-stringops
8720 @opindex minline-all-stringops
8721 By default GCC inlines string operations only when destination is known to be
8722 aligned at least to 4 byte boundary. This enables more inlining, increase code
8723 size, but may improve performance of code that depends on fast memcpy, strlen
8724 and memset for short lengths.
8726 @item -momit-leaf-frame-pointer
8727 @opindex momit-leaf-frame-pointer
8728 Don't keep the frame pointer in a register for leaf functions. This
8729 avoids the instructions to save, set up and restore frame pointers and
8730 makes an extra register available in leaf functions. The option
8731 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8732 which might make debugging harder.
8734 @item -mtls-direct-seg-refs
8735 @itemx -mno-tls-direct-seg-refs
8736 @opindex mtls-direct-seg-refs
8737 Controls whether TLS variables may be accessed with offsets from the
8738 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
8739 or whether the thread base pointer must be added. Whether or not this
8740 is legal depends on the operating system, and whether it maps the
8741 segment to cover the entire TLS area.
8743 For systems that use GNU libc, the default is on.
8746 These @samp{-m} switches are supported in addition to the above
8747 on AMD x86-64 processors in 64-bit environments.
8754 Generate code for a 32-bit or 64-bit environment.
8755 The 32-bit environment sets int, long and pointer to 32 bits and
8756 generates code that runs on any i386 system.
8757 The 64-bit environment sets int to 32 bits and long and pointer
8758 to 64 bits and generates code for AMD's x86-64 architecture.
8761 @opindex no-red-zone
8762 Do not use a so called red zone for x86-64 code. The red zone is mandated
8763 by the x86-64 ABI, it is a 128-byte area beyond the location of the
8764 stack pointer that will not be modified by signal or interrupt handlers
8765 and therefore can be used for temporary data without adjusting the stack
8766 pointer. The flag @option{-mno-red-zone} disables this red zone.
8768 @item -mcmodel=small
8769 @opindex mcmodel=small
8770 Generate code for the small code model: the program and its symbols must
8771 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
8772 Programs can be statically or dynamically linked. This is the default
8775 @item -mcmodel=kernel
8776 @opindex mcmodel=kernel
8777 Generate code for the kernel code model. The kernel runs in the
8778 negative 2 GB of the address space.
8779 This model has to be used for Linux kernel code.
8781 @item -mcmodel=medium
8782 @opindex mcmodel=medium
8783 Generate code for the medium model: The program is linked in the lower 2
8784 GB of the address space but symbols can be located anywhere in the
8785 address space. Programs can be statically or dynamically linked, but
8786 building of shared libraries are not supported with the medium model.
8788 @item -mcmodel=large
8789 @opindex mcmodel=large
8790 Generate code for the large model: This model makes no assumptions
8791 about addresses and sizes of sections. Currently GCC does not implement
8796 @subsection IA-64 Options
8797 @cindex IA-64 Options
8799 These are the @samp{-m} options defined for the Intel IA-64 architecture.
8803 @opindex mbig-endian
8804 Generate code for a big endian target. This is the default for HP-UX@.
8806 @item -mlittle-endian
8807 @opindex mlittle-endian
8808 Generate code for a little endian target. This is the default for AIX5
8815 Generate (or don't) code for the GNU assembler. This is the default.
8816 @c Also, this is the default if the configure option @option{--with-gnu-as}
8823 Generate (or don't) code for the GNU linker. This is the default.
8824 @c Also, this is the default if the configure option @option{--with-gnu-ld}
8829 Generate code that does not use a global pointer register. The result
8830 is not position independent code, and violates the IA-64 ABI@.
8832 @item -mvolatile-asm-stop
8833 @itemx -mno-volatile-asm-stop
8834 @opindex mvolatile-asm-stop
8835 @opindex mno-volatile-asm-stop
8836 Generate (or don't) a stop bit immediately before and after volatile asm
8841 Generate code that works around Itanium B step errata.
8843 @item -mregister-names
8844 @itemx -mno-register-names
8845 @opindex mregister-names
8846 @opindex mno-register-names
8847 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
8848 the stacked registers. This may make assembler output more readable.
8854 Disable (or enable) optimizations that use the small data section. This may
8855 be useful for working around optimizer bugs.
8858 @opindex mconstant-gp
8859 Generate code that uses a single constant global pointer value. This is
8860 useful when compiling kernel code.
8864 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
8865 This is useful when compiling firmware code.
8867 @item -minline-float-divide-min-latency
8868 @opindex minline-float-divide-min-latency
8869 Generate code for inline divides of floating point values
8870 using the minimum latency algorithm.
8872 @item -minline-float-divide-max-throughput
8873 @opindex minline-float-divide-max-throughput
8874 Generate code for inline divides of floating point values
8875 using the maximum throughput algorithm.
8877 @item -minline-int-divide-min-latency
8878 @opindex minline-int-divide-min-latency
8879 Generate code for inline divides of integer values
8880 using the minimum latency algorithm.
8882 @item -minline-int-divide-max-throughput
8883 @opindex minline-int-divide-max-throughput
8884 Generate code for inline divides of integer values
8885 using the maximum throughput algorithm.
8887 @item -minline-sqrt-min-latency
8888 @opindex minline-sqrt-min-latency
8889 Generate code for inline square roots
8890 using the minimum latency algorithm.
8892 @item -minline-sqrt-max-throughput
8893 @opindex minline-sqrt-max-throughput
8894 Generate code for inline square roots
8895 using the maximum throughput algorithm.
8897 @item -mno-dwarf2-asm
8899 @opindex mno-dwarf2-asm
8900 @opindex mdwarf2-asm
8901 Don't (or do) generate assembler code for the DWARF2 line number debugging
8902 info. This may be useful when not using the GNU assembler.
8904 @item -mearly-stop-bits
8905 @itemx -mno-early-stop-bits
8906 @opindex mearly-stop-bits
8907 @opindex mno-early-stop-bits
8908 Allow stop bits to be placed earlier than immediately preceding the
8909 instruction that triggered the stop bit. This can improve instruction
8910 scheduling, but does not always do so.
8912 @item -mfixed-range=@var{register-range}
8913 @opindex mfixed-range
8914 Generate code treating the given register range as fixed registers.
8915 A fixed register is one that the register allocator can not use. This is
8916 useful when compiling kernel code. A register range is specified as
8917 two registers separated by a dash. Multiple register ranges can be
8918 specified separated by a comma.
8920 @item -mtls-size=@var{tls-size}
8922 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
8925 @item -mtune-arch=@var{cpu-type}
8927 Tune the instruction scheduling for a particular CPU, Valid values are
8928 itanium, itanium1, merced, itanium2, and mckinley.
8934 Add support for multithreading using the POSIX threads library. This
8935 option sets flags for both the preprocessor and linker. It does
8936 not affect the thread safety of object code produced by the compiler or
8937 that of libraries supplied with it. These are HP-UX specific flags.
8943 Generate code for a 32-bit or 64-bit environment.
8944 The 32-bit environment sets int, long and pointer to 32 bits.
8945 The 64-bit environment sets int to 32 bits and long and pointer
8946 to 64 bits. These are HP-UX specific flags.
8950 @node M32R/D Options
8951 @subsection M32R/D Options
8952 @cindex M32R/D options
8954 These @option{-m} options are defined for Renesas M32R/D architectures:
8959 Generate code for the M32R/2@.
8963 Generate code for the M32R/X@.
8967 Generate code for the M32R@. This is the default.
8970 @opindex mmodel=small
8971 Assume all objects live in the lower 16MB of memory (so that their addresses
8972 can be loaded with the @code{ld24} instruction), and assume all subroutines
8973 are reachable with the @code{bl} instruction.
8974 This is the default.
8976 The addressability of a particular object can be set with the
8977 @code{model} attribute.
8979 @item -mmodel=medium
8980 @opindex mmodel=medium
8981 Assume objects may be anywhere in the 32-bit address space (the compiler
8982 will generate @code{seth/add3} instructions to load their addresses), and
8983 assume all subroutines are reachable with the @code{bl} instruction.
8986 @opindex mmodel=large
8987 Assume objects may be anywhere in the 32-bit address space (the compiler
8988 will generate @code{seth/add3} instructions to load their addresses), and
8989 assume subroutines may not be reachable with the @code{bl} instruction
8990 (the compiler will generate the much slower @code{seth/add3/jl}
8991 instruction sequence).
8994 @opindex msdata=none
8995 Disable use of the small data area. Variables will be put into
8996 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
8997 @code{section} attribute has been specified).
8998 This is the default.
9000 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
9001 Objects may be explicitly put in the small data area with the
9002 @code{section} attribute using one of these sections.
9005 @opindex msdata=sdata
9006 Put small global and static data in the small data area, but do not
9007 generate special code to reference them.
9011 Put small global and static data in the small data area, and generate
9012 special instructions to reference them.
9016 @cindex smaller data references
9017 Put global and static objects less than or equal to @var{num} bytes
9018 into the small data or bss sections instead of the normal data or bss
9019 sections. The default value of @var{num} is 8.
9020 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
9021 for this option to have any effect.
9023 All modules should be compiled with the same @option{-G @var{num}} value.
9024 Compiling with different values of @var{num} may or may not work; if it
9025 doesn't the linker will give an error message---incorrect code will not be
9030 Makes the M32R specific code in the compiler display some statistics
9031 that might help in debugging programs.
9034 @opindex malign-loops
9035 Align all loops to a 32-byte boundary.
9037 @item -mno-align-loops
9038 @opindex mno-align-loops
9039 Do not enforce a 32-byte alignment for loops. This is the default.
9041 @item -missue-rate=@var{number}
9042 @opindex missue-rate=@var{number}
9043 Issue @var{number} instructions per cycle. @var{number} can only be 1
9046 @item -mbranch-cost=@var{number}
9047 @opindex mbranch-cost=@var{number}
9048 @var{number} can only be 1 or 2. If it is 1 then branches will be
9049 preferred over conditional code, if it is 2, then the opposite will
9052 @item -mflush-trap=@var{number}
9053 @opindex mflush-trap=@var{number}
9054 Specifies the trap number to use to flush the cache. The default is
9055 12. Valid numbers are between 0 and 15 inclusive.
9057 @item -mno-flush-trap
9058 @opindex mno-flush-trap
9059 Specifies that the cache cannot be flushed by using a trap.
9061 @item -mflush-func=@var{name}
9062 @opindex mflush-func=@var{name}
9063 Specifies the name of the operating system function to call to flush
9064 the cache. The default is @emph{_flush_cache}, but a function call
9065 will only be used if a trap is not available.
9067 @item -mno-flush-func
9068 @opindex mno-flush-func
9069 Indicates that there is no OS function for flushing the cache.
9073 @node M680x0 Options
9074 @subsection M680x0 Options
9075 @cindex M680x0 options
9077 These are the @samp{-m} options defined for the 68000 series. The default
9078 values for these options depends on which style of 68000 was selected when
9079 the compiler was configured; the defaults for the most common choices are
9087 Generate output for a 68000. This is the default
9088 when the compiler is configured for 68000-based systems.
9090 Use this option for microcontrollers with a 68000 or EC000 core,
9091 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
9097 Generate output for a 68020. This is the default
9098 when the compiler is configured for 68020-based systems.
9102 Generate output containing 68881 instructions for floating point.
9103 This is the default for most 68020 systems unless @option{--nfp} was
9104 specified when the compiler was configured.
9108 Generate output for a 68030. This is the default when the compiler is
9109 configured for 68030-based systems.
9113 Generate output for a 68040. This is the default when the compiler is
9114 configured for 68040-based systems.
9116 This option inhibits the use of 68881/68882 instructions that have to be
9117 emulated by software on the 68040. Use this option if your 68040 does not
9118 have code to emulate those instructions.
9122 Generate output for a 68060. This is the default when the compiler is
9123 configured for 68060-based systems.
9125 This option inhibits the use of 68020 and 68881/68882 instructions that
9126 have to be emulated by software on the 68060. Use this option if your 68060
9127 does not have code to emulate those instructions.
9131 Generate output for a CPU32. This is the default
9132 when the compiler is configured for CPU32-based systems.
9134 Use this option for microcontrollers with a
9135 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
9136 68336, 68340, 68341, 68349 and 68360.
9140 Generate output for a 520X ``coldfire'' family cpu. This is the default
9141 when the compiler is configured for 520X-based systems.
9143 Use this option for microcontroller with a 5200 core, including
9144 the MCF5202, MCF5203, MCF5204 and MCF5202.
9149 Generate output for a 68040, without using any of the new instructions.
9150 This results in code which can run relatively efficiently on either a
9151 68020/68881 or a 68030 or a 68040. The generated code does use the
9152 68881 instructions that are emulated on the 68040.
9156 Generate output for a 68060, without using any of the new instructions.
9157 This results in code which can run relatively efficiently on either a
9158 68020/68881 or a 68030 or a 68040. The generated code does use the
9159 68881 instructions that are emulated on the 68060.
9162 @opindex msoft-float
9163 Generate output containing library calls for floating point.
9164 @strong{Warning:} the requisite libraries are not available for all m68k
9165 targets. Normally the facilities of the machine's usual C compiler are
9166 used, but this can't be done directly in cross-compilation. You must
9167 make your own arrangements to provide suitable library functions for
9168 cross-compilation. The embedded targets @samp{m68k-*-aout} and
9169 @samp{m68k-*-coff} do provide software floating point support.
9173 Consider type @code{int} to be 16 bits wide, like @code{short int}.
9174 Additionally, parameters passed on the stack are also aligned to a
9175 16-bit boundary even on targets whose API mandates promotion to 32-bit.
9178 @opindex mnobitfield
9179 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
9180 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
9184 Do use the bit-field instructions. The @option{-m68020} option implies
9185 @option{-mbitfield}. This is the default if you use a configuration
9186 designed for a 68020.
9190 Use a different function-calling convention, in which functions
9191 that take a fixed number of arguments return with the @code{rtd}
9192 instruction, which pops their arguments while returning. This
9193 saves one instruction in the caller since there is no need to pop
9194 the arguments there.
9196 This calling convention is incompatible with the one normally
9197 used on Unix, so you cannot use it if you need to call libraries
9198 compiled with the Unix compiler.
9200 Also, you must provide function prototypes for all functions that
9201 take variable numbers of arguments (including @code{printf});
9202 otherwise incorrect code will be generated for calls to those
9205 In addition, seriously incorrect code will result if you call a
9206 function with too many arguments. (Normally, extra arguments are
9207 harmlessly ignored.)
9209 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
9210 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
9213 @itemx -mno-align-int
9215 @opindex mno-align-int
9216 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
9217 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
9218 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
9219 Aligning variables on 32-bit boundaries produces code that runs somewhat
9220 faster on processors with 32-bit busses at the expense of more memory.
9222 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
9223 align structures containing the above types differently than
9224 most published application binary interface specifications for the m68k.
9228 Use the pc-relative addressing mode of the 68000 directly, instead of
9229 using a global offset table. At present, this option implies @option{-fpic},
9230 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
9231 not presently supported with @option{-mpcrel}, though this could be supported for
9232 68020 and higher processors.
9234 @item -mno-strict-align
9235 @itemx -mstrict-align
9236 @opindex mno-strict-align
9237 @opindex mstrict-align
9238 Do not (do) assume that unaligned memory references will be handled by
9242 Generate code that allows the data segment to be located in a different
9243 area of memory from the text segment. This allows for execute in place in
9244 an environment without virtual memory management. This option implies
9248 Generate code that assumes that the data segment follows the text segment.
9249 This is the default.
9251 @item -mid-shared-library
9252 Generate code that supports shared libraries via the library ID method.
9253 This allows for execute in place and shared libraries in an environment
9254 without virtual memory management. This option implies @option{-fPIC}.
9256 @item -mno-id-shared-library
9257 Generate code that doesn't assume ID based shared libraries are being used.
9258 This is the default.
9260 @item -mshared-library-id=n
9261 Specified the identification number of the ID based shared library being
9262 compiled. Specifying a value of 0 will generate more compact code, specifying
9263 other values will force the allocation of that number to the current
9264 library but is no more space or time efficient than omitting this option.
9268 @node M68hc1x Options
9269 @subsection M68hc1x Options
9270 @cindex M68hc1x options
9272 These are the @samp{-m} options defined for the 68hc11 and 68hc12
9273 microcontrollers. The default values for these options depends on
9274 which style of microcontroller was selected when the compiler was configured;
9275 the defaults for the most common choices are given below.
9282 Generate output for a 68HC11. This is the default
9283 when the compiler is configured for 68HC11-based systems.
9289 Generate output for a 68HC12. This is the default
9290 when the compiler is configured for 68HC12-based systems.
9296 Generate output for a 68HCS12.
9299 @opindex mauto-incdec
9300 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
9307 Enable the use of 68HC12 min and max instructions.
9310 @itemx -mno-long-calls
9311 @opindex mlong-calls
9312 @opindex mno-long-calls
9313 Treat all calls as being far away (near). If calls are assumed to be
9314 far away, the compiler will use the @code{call} instruction to
9315 call a function and the @code{rtc} instruction for returning.
9319 Consider type @code{int} to be 16 bits wide, like @code{short int}.
9321 @item -msoft-reg-count=@var{count}
9322 @opindex msoft-reg-count
9323 Specify the number of pseudo-soft registers which are used for the
9324 code generation. The maximum number is 32. Using more pseudo-soft
9325 register may or may not result in better code depending on the program.
9326 The default is 4 for 68HC11 and 2 for 68HC12.
9331 @subsection MCore Options
9332 @cindex MCore options
9334 These are the @samp{-m} options defined for the Motorola M*Core
9342 @opindex mno-hardlit
9343 Inline constants into the code stream if it can be done in two
9344 instructions or less.
9350 Use the divide instruction. (Enabled by default).
9352 @item -mrelax-immediate
9353 @itemx -mno-relax-immediate
9354 @opindex mrelax-immediate
9355 @opindex mno-relax-immediate
9356 Allow arbitrary sized immediates in bit operations.
9358 @item -mwide-bitfields
9359 @itemx -mno-wide-bitfields
9360 @opindex mwide-bitfields
9361 @opindex mno-wide-bitfields
9362 Always treat bit-fields as int-sized.
9364 @item -m4byte-functions
9365 @itemx -mno-4byte-functions
9366 @opindex m4byte-functions
9367 @opindex mno-4byte-functions
9368 Force all functions to be aligned to a four byte boundary.
9370 @item -mcallgraph-data
9371 @itemx -mno-callgraph-data
9372 @opindex mcallgraph-data
9373 @opindex mno-callgraph-data
9374 Emit callgraph information.
9377 @itemx -mno-slow-bytes
9378 @opindex mslow-bytes
9379 @opindex mno-slow-bytes
9380 Prefer word access when reading byte quantities.
9382 @item -mlittle-endian
9384 @opindex mlittle-endian
9385 @opindex mbig-endian
9386 Generate code for a little endian target.
9392 Generate code for the 210 processor.
9396 @subsection MIPS Options
9397 @cindex MIPS options
9403 Generate big-endian code.
9407 Generate little-endian code. This is the default for @samp{mips*el-*-*}
9410 @item -march=@var{arch}
9412 Generate code that will run on @var{arch}, which can be the name of a
9413 generic MIPS ISA, or the name of a particular processor.
9415 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
9416 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
9417 The processor names are:
9418 @samp{4kc}, @samp{4kp}, @samp{5kc}, @samp{20kc},
9420 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
9421 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000}, @samp{rm7000},
9425 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
9426 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
9427 The special value @samp{from-abi} selects the
9428 most compatible architecture for the selected ABI (that is,
9429 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
9431 In processor names, a final @samp{000} can be abbreviated as @samp{k}
9432 (for example, @samp{-march=r2k}). Prefixes are optional, and
9433 @samp{vr} may be written @samp{r}.
9435 GCC defines two macros based on the value of this option. The first
9436 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
9437 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
9438 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
9439 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
9440 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
9442 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
9443 above. In other words, it will have the full prefix and will not
9444 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
9445 the macro names the resolved architecture (either @samp{"mips1"} or
9446 @samp{"mips3"}). It names the default architecture when no
9447 @option{-march} option is given.
9449 @item -mtune=@var{arch}
9451 Optimize for @var{arch}. Among other things, this option controls
9452 the way instructions are scheduled, and the perceived cost of arithmetic
9453 operations. The list of @var{arch} values is the same as for
9456 When this option is not used, GCC will optimize for the processor
9457 specified by @option{-march}. By using @option{-march} and
9458 @option{-mtune} together, it is possible to generate code that will
9459 run on a family of processors, but optimize the code for one
9460 particular member of that family.
9462 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
9463 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
9464 @samp{-march} ones described above.
9468 Equivalent to @samp{-march=mips1}.
9472 Equivalent to @samp{-march=mips2}.
9476 Equivalent to @samp{-march=mips3}.
9480 Equivalent to @samp{-march=mips4}.
9484 Equivalent to @samp{-march=mips32}.
9488 Equivalent to @samp{-march=mips32r2}.
9492 Equivalent to @samp{-march=mips64}.
9498 Use (do not use) the MIPS16 ISA@.
9510 Generate code for the given ABI@.
9512 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
9513 generates 64-bit code when you select a 64-bit architecture, but you
9514 can use @option{-mgp32} to get 32-bit code instead.
9516 For information about the O64 ABI, see
9517 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
9520 @itemx -mno-abicalls
9522 @opindex mno-abicalls
9523 Generate (do not generate) SVR4-style position-independent code.
9524 @option{-mabicalls} is the default for SVR4-based systems.
9530 Lift (do not lift) the usual restrictions on the size of the global
9533 GCC normally uses a single instruction to load values from the GOT@.
9534 While this is relatively efficient, it will only work if the GOT
9535 is smaller than about 64k. Anything larger will cause the linker
9536 to report an error such as:
9538 @cindex relocation truncated to fit (MIPS)
9540 relocation truncated to fit: R_MIPS_GOT16 foobar
9543 If this happens, you should recompile your code with @option{-mxgot}.
9544 It should then work with very large GOTs, although it will also be
9545 less efficient, since it will take three instructions to fetch the
9546 value of a global symbol.
9548 Note that some linkers can create multiple GOTs. If you have such a
9549 linker, you should only need to use @option{-mxgot} when a single object
9550 file accesses more than 64k's worth of GOT entries. Very few do.
9552 These options have no effect unless GCC is generating position
9557 Assume that general-purpose registers are 32 bits wide.
9561 Assume that general-purpose registers are 64 bits wide.
9565 Assume that floating-point registers are 32 bits wide.
9569 Assume that floating-point registers are 64 bits wide.
9572 @opindex mhard-float
9573 Use floating-point coprocessor instructions.
9576 @opindex msoft-float
9577 Do not use floating-point coprocessor instructions. Implement
9578 floating-point calculations using library calls instead.
9580 @item -msingle-float
9581 @opindex msingle-float
9582 Assume that the floating-point coprocessor only supports single-precision
9585 @itemx -mdouble-float
9586 @opindex mdouble-float
9587 Assume that the floating-point coprocessor supports double-precision
9588 operations. This is the default.
9590 @itemx -mpaired-single
9591 @itemx -mno-paired-single
9592 @opindex mpaired-single
9593 @opindex mno-paired-single
9594 Use (do not use) paired-single floating-point instructions.
9595 @xref{MIPS Paired-Single Support}. This option can only be used
9596 when generating 64-bit code and requires hardware floating-point
9597 support to be enabled.
9603 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
9604 The option @option{-mips3d} implies @option{-mpaired-single}.
9608 Force @code{int} and @code{long} types to be 64 bits wide. See
9609 @option{-mlong32} for an explanation of the default and the way
9610 that the pointer size is determined.
9612 This option has been deprecated and will be removed in a future release.
9616 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
9617 an explanation of the default and the way that the pointer size is
9622 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
9624 The default size of @code{int}s, @code{long}s and pointers depends on
9625 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
9626 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
9627 32-bit @code{long}s. Pointers are the same size as @code{long}s,
9628 or the same size as integer registers, whichever is smaller.
9632 @cindex smaller data references (MIPS)
9633 @cindex gp-relative references (MIPS)
9634 Put global and static items less than or equal to @var{num} bytes into
9635 the small data or bss section instead of the normal data or bss section.
9636 This allows the data to be accessed using a single instruction.
9638 All modules should be compiled with the same @option{-G @var{num}}
9641 @item -membedded-data
9642 @itemx -mno-embedded-data
9643 @opindex membedded-data
9644 @opindex mno-embedded-data
9645 Allocate variables to the read-only data section first if possible, then
9646 next in the small data section if possible, otherwise in data. This gives
9647 slightly slower code than the default, but reduces the amount of RAM required
9648 when executing, and thus may be preferred for some embedded systems.
9650 @item -muninit-const-in-rodata
9651 @itemx -mno-uninit-const-in-rodata
9652 @opindex muninit-const-in-rodata
9653 @opindex mno-uninit-const-in-rodata
9654 Put uninitialized @code{const} variables in the read-only data section.
9655 This option is only meaningful in conjunction with @option{-membedded-data}.
9657 @item -msplit-addresses
9658 @itemx -mno-split-addresses
9659 @opindex msplit-addresses
9660 @opindex mno-split-addresses
9661 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
9662 relocation operators. This option has been superceded by
9663 @option{-mexplicit-relocs} but is retained for backwards compatibility.
9665 @item -mexplicit-relocs
9666 @itemx -mno-explicit-relocs
9667 @opindex mexplicit-relocs
9668 @opindex mno-explicit-relocs
9669 Use (do not use) assembler relocation operators when dealing with symbolic
9670 addresses. The alternative, selected by @option{-mno-explicit-relocs},
9671 is to use assembler macros instead.
9673 @option{-mexplicit-relocs} is the default if GCC was configured
9674 to use an assembler that supports relocation operators.
9676 @item -mcheck-zero-division
9677 @itemx -mno-check-zero-division
9678 @opindex mcheck-zero-division
9679 @opindex mno-check-zero-division
9680 Trap (do not trap) on integer division by zero. The default is
9681 @option{-mcheck-zero-division}.
9683 @item -mdivide-traps
9684 @itemx -mdivide-breaks
9685 @opindex mdivide-traps
9686 @opindex mdivide-breaks
9687 MIPS systems check for division by zero by generating either a
9688 conditional trap or a break instruction. Using traps results in
9689 smaller code, but is only supported on MIPS II and later. Also, some
9690 versions of the Linux kernel have a bug that prevents trap from
9691 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
9692 allow conditional traps on architectures that support them and
9693 @option{-mdivide-breaks} to force the use of breaks.
9695 The default is usually @option{-mdivide-traps}, but this can be
9696 overridden at configure time using @option{--with-divide=breaks}.
9697 Divide-by-zero checks can be completely disabled using
9698 @option{-mno-check-zero-division}.
9704 Force (do not force) the use of @code{memcpy()} for non-trivial block
9705 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
9706 most constant-sized copies.
9709 @itemx -mno-long-calls
9710 @opindex mlong-calls
9711 @opindex mno-long-calls
9712 Disable (do not disable) use of the @code{jal} instruction. Calling
9713 functions using @code{jal} is more efficient but requires the caller
9714 and callee to be in the same 256 megabyte segment.
9716 This option has no effect on abicalls code. The default is
9717 @option{-mno-long-calls}.
9723 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
9724 instructions, as provided by the R4650 ISA@.
9727 @itemx -mno-fused-madd
9728 @opindex mfused-madd
9729 @opindex mno-fused-madd
9730 Enable (disable) use of the floating point multiply-accumulate
9731 instructions, when they are available. The default is
9732 @option{-mfused-madd}.
9734 When multiply-accumulate instructions are used, the intermediate
9735 product is calculated to infinite precision and is not subject to
9736 the FCSR Flush to Zero bit. This may be undesirable in some
9741 Tell the MIPS assembler to not run its preprocessor over user
9742 assembler files (with a @samp{.s} suffix) when assembling them.
9745 @itemx -mno-fix-r4000
9747 @opindex mno-fix-r4000
9748 Work around certain R4000 CPU errata:
9751 A double-word or a variable shift may give an incorrect result if executed
9752 immediately after starting an integer division.
9754 A double-word or a variable shift may give an incorrect result if executed
9755 while an integer multiplication is in progress.
9757 An integer division may give an incorrect result if started in a delay slot
9758 of a taken branch or a jump.
9762 @itemx -mno-fix-r4400
9764 @opindex mno-fix-r4400
9765 Work around certain R4400 CPU errata:
9768 A double-word or a variable shift may give an incorrect result if executed
9769 immediately after starting an integer division.
9773 @itemx -mno-fix-vr4120
9774 @opindex mfix-vr4120
9775 Work around certain VR4120 errata:
9778 @code{dmultu} does not always produce the correct result.
9780 @code{div} and @code{ddiv} do not always produce the correct result if one
9781 of the operands is negative.
9783 The workarounds for the division errata rely on special functions in
9784 @file{libgcc.a}. At present, these functions are only provided by
9785 the @code{mips64vr*-elf} configurations.
9787 Other VR4120 errata require a nop to be inserted between certain pairs of
9788 instructions. These errata are handled by the assembler, not by GCC itself.
9793 Work around certain SB-1 CPU core errata.
9794 (This flag currently works around the SB-1 revision 2
9795 ``F1'' and ``F2'' floating point errata.)
9797 @item -mflush-func=@var{func}
9798 @itemx -mno-flush-func
9799 @opindex mflush-func
9800 Specifies the function to call to flush the I and D caches, or to not
9801 call any such function. If called, the function must take the same
9802 arguments as the common @code{_flush_func()}, that is, the address of the
9803 memory range for which the cache is being flushed, the size of the
9804 memory range, and the number 3 (to flush both caches). The default
9805 depends on the target GCC was configured for, but commonly is either
9806 @samp{_flush_func} or @samp{__cpu_flush}.
9808 @item -mbranch-likely
9809 @itemx -mno-branch-likely
9810 @opindex mbranch-likely
9811 @opindex mno-branch-likely
9812 Enable or disable use of Branch Likely instructions, regardless of the
9813 default for the selected architecture. By default, Branch Likely
9814 instructions may be generated if they are supported by the selected
9815 architecture. An exception is for the MIPS32 and MIPS64 architectures
9816 and processors which implement those architectures; for those, Branch
9817 Likely instructions will not be generated by default because the MIPS32
9818 and MIPS64 architectures specifically deprecate their use.
9820 @item -mfp-exceptions
9821 @itemx -mno-fp-exceptions
9822 @opindex mfp-exceptions
9823 Specifies whether FP exceptions are enabled. This affects how we schedule
9824 FP instructions for some processors. The default is that FP exceptions are
9827 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
9828 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
9831 @item -mvr4130-align
9832 @itemx -mno-vr4130-align
9833 @opindex mvr4130-align
9834 The VR4130 pipeline is two-way superscalar, but can only issue two
9835 instructions together if the first one is 8-byte aligned. When this
9836 option is enabled, GCC will align pairs of instructions that it
9837 thinks should execute in parallel.
9839 This option only has an effect when optimizing for the VR4130.
9840 It normally makes code faster, but at the expense of making it bigger.
9841 It is enabled by default at optimization level @option{-O3}.
9845 @subsection MMIX Options
9846 @cindex MMIX Options
9848 These options are defined for the MMIX:
9852 @itemx -mno-libfuncs
9854 @opindex mno-libfuncs
9855 Specify that intrinsic library functions are being compiled, passing all
9856 values in registers, no matter the size.
9861 @opindex mno-epsilon
9862 Generate floating-point comparison instructions that compare with respect
9863 to the @code{rE} epsilon register.
9865 @item -mabi=mmixware
9867 @opindex mabi-mmixware
9869 Generate code that passes function parameters and return values that (in
9870 the called function) are seen as registers @code{$0} and up, as opposed to
9871 the GNU ABI which uses global registers @code{$231} and up.
9874 @itemx -mno-zero-extend
9875 @opindex mzero-extend
9876 @opindex mno-zero-extend
9877 When reading data from memory in sizes shorter than 64 bits, use (do not
9878 use) zero-extending load instructions by default, rather than
9879 sign-extending ones.
9882 @itemx -mno-knuthdiv
9884 @opindex mno-knuthdiv
9885 Make the result of a division yielding a remainder have the same sign as
9886 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9887 remainder follows the sign of the dividend. Both methods are
9888 arithmetically valid, the latter being almost exclusively used.
9890 @item -mtoplevel-symbols
9891 @itemx -mno-toplevel-symbols
9892 @opindex mtoplevel-symbols
9893 @opindex mno-toplevel-symbols
9894 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9895 code can be used with the @code{PREFIX} assembly directive.
9899 Generate an executable in the ELF format, rather than the default
9900 @samp{mmo} format used by the @command{mmix} simulator.
9902 @item -mbranch-predict
9903 @itemx -mno-branch-predict
9904 @opindex mbranch-predict
9905 @opindex mno-branch-predict
9906 Use (do not use) the probable-branch instructions, when static branch
9907 prediction indicates a probable branch.
9909 @item -mbase-addresses
9910 @itemx -mno-base-addresses
9911 @opindex mbase-addresses
9912 @opindex mno-base-addresses
9913 Generate (do not generate) code that uses @emph{base addresses}. Using a
9914 base address automatically generates a request (handled by the assembler
9915 and the linker) for a constant to be set up in a global register. The
9916 register is used for one or more base address requests within the range 0
9917 to 255 from the value held in the register. The generally leads to short
9918 and fast code, but the number of different data items that can be
9919 addressed is limited. This means that a program that uses lots of static
9920 data may require @option{-mno-base-addresses}.
9923 @itemx -mno-single-exit
9924 @opindex msingle-exit
9925 @opindex mno-single-exit
9926 Force (do not force) generated code to have a single exit point in each
9930 @node MN10300 Options
9931 @subsection MN10300 Options
9932 @cindex MN10300 options
9934 These @option{-m} options are defined for Matsushita MN10300 architectures:
9939 Generate code to avoid bugs in the multiply instructions for the MN10300
9940 processors. This is the default.
9943 @opindex mno-mult-bug
9944 Do not generate code to avoid bugs in the multiply instructions for the
9949 Generate code which uses features specific to the AM33 processor.
9953 Do not generate code which uses features specific to the AM33 processor. This
9958 Do not link in the C run-time initialization object file.
9962 Indicate to the linker that it should perform a relaxation optimization pass
9963 to shorten branches, calls and absolute memory addresses. This option only
9964 has an effect when used on the command line for the final link step.
9966 This option makes symbolic debugging impossible.
9970 @subsection NS32K Options
9971 @cindex NS32K options
9973 These are the @samp{-m} options defined for the 32000 series. The default
9974 values for these options depends on which style of 32000 was selected when
9975 the compiler was configured; the defaults for the most common choices are
9983 Generate output for a 32032. This is the default
9984 when the compiler is configured for 32032 and 32016 based systems.
9990 Generate output for a 32332. This is the default
9991 when the compiler is configured for 32332-based systems.
9997 Generate output for a 32532. This is the default
9998 when the compiler is configured for 32532-based systems.
10002 Generate output containing 32081 instructions for floating point.
10003 This is the default for all systems.
10007 Generate output containing 32381 instructions for floating point. This
10008 also implies @option{-m32081}. The 32381 is only compatible with the 32332
10009 and 32532 cpus. This is the default for the pc532-netbsd configuration.
10012 @opindex mmulti-add
10013 Try and generate multiply-add floating point instructions @code{polyF}
10014 and @code{dotF}. This option is only available if the @option{-m32381}
10015 option is in effect. Using these instructions requires changes to
10016 register allocation which generally has a negative impact on
10017 performance. This option should only be enabled when compiling code
10018 particularly likely to make heavy use of multiply-add instructions.
10020 @item -mnomulti-add
10021 @opindex mnomulti-add
10022 Do not try and generate multiply-add floating point instructions
10023 @code{polyF} and @code{dotF}. This is the default on all platforms.
10026 @opindex msoft-float
10027 Generate output containing library calls for floating point.
10028 @strong{Warning:} the requisite libraries may not be available.
10030 @item -mieee-compare
10031 @itemx -mno-ieee-compare
10032 @opindex mieee-compare
10033 @opindex mno-ieee-compare
10034 Control whether or not the compiler uses IEEE floating point
10035 comparisons. These handle correctly the case where the result of a
10036 comparison is unordered.
10037 @strong{Warning:} the requisite kernel support may not be available.
10040 @opindex mnobitfield
10041 Do not use the bit-field instructions. On some machines it is faster to
10042 use shifting and masking operations. This is the default for the pc532.
10046 Do use the bit-field instructions. This is the default for all platforms
10051 Use a different function-calling convention, in which functions
10052 that take a fixed number of arguments return pop their
10053 arguments on return with the @code{ret} instruction.
10055 This calling convention is incompatible with the one normally
10056 used on Unix, so you cannot use it if you need to call libraries
10057 compiled with the Unix compiler.
10059 Also, you must provide function prototypes for all functions that
10060 take variable numbers of arguments (including @code{printf});
10061 otherwise incorrect code will be generated for calls to those
10064 In addition, seriously incorrect code will result if you call a
10065 function with too many arguments. (Normally, extra arguments are
10066 harmlessly ignored.)
10068 This option takes its name from the 680x0 @code{rtd} instruction.
10073 Use a different function-calling convention where the first two arguments
10074 are passed in registers.
10076 This calling convention is incompatible with the one normally
10077 used on Unix, so you cannot use it if you need to call libraries
10078 compiled with the Unix compiler.
10081 @opindex mnoregparam
10082 Do not pass any arguments in registers. This is the default for all
10087 It is OK to use the sb as an index register which is always loaded with
10088 zero. This is the default for the pc532-netbsd target.
10092 The sb register is not available for use or has not been initialized to
10093 zero by the run time system. This is the default for all targets except
10094 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
10095 @option{-fpic} is set.
10099 Many ns32000 series addressing modes use displacements of up to 512MB@.
10100 If an address is above 512MB then displacements from zero can not be used.
10101 This option causes code to be generated which can be loaded above 512MB@.
10102 This may be useful for operating systems or ROM code.
10106 Assume code will be loaded in the first 512MB of virtual address space.
10107 This is the default for all platforms.
10111 @node PDP-11 Options
10112 @subsection PDP-11 Options
10113 @cindex PDP-11 Options
10115 These options are defined for the PDP-11:
10120 Use hardware FPP floating point. This is the default. (FIS floating
10121 point on the PDP-11/40 is not supported.)
10124 @opindex msoft-float
10125 Do not use hardware floating point.
10129 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
10133 Return floating-point results in memory. This is the default.
10137 Generate code for a PDP-11/40.
10141 Generate code for a PDP-11/45. This is the default.
10145 Generate code for a PDP-11/10.
10147 @item -mbcopy-builtin
10148 @opindex bcopy-builtin
10149 Use inline @code{movmemhi} patterns for copying memory. This is the
10154 Do not use inline @code{movmemhi} patterns for copying memory.
10160 Use 16-bit @code{int}. This is the default.
10166 Use 32-bit @code{int}.
10169 @itemx -mno-float32
10171 @opindex mno-float32
10172 Use 64-bit @code{float}. This is the default.
10175 @itemx -mno-float64
10177 @opindex mno-float64
10178 Use 32-bit @code{float}.
10182 Use @code{abshi2} pattern. This is the default.
10186 Do not use @code{abshi2} pattern.
10188 @item -mbranch-expensive
10189 @opindex mbranch-expensive
10190 Pretend that branches are expensive. This is for experimenting with
10191 code generation only.
10193 @item -mbranch-cheap
10194 @opindex mbranch-cheap
10195 Do not pretend that branches are expensive. This is the default.
10199 Generate code for a system with split I&D@.
10203 Generate code for a system without split I&D@. This is the default.
10207 Use Unix assembler syntax. This is the default when configured for
10208 @samp{pdp11-*-bsd}.
10212 Use DEC assembler syntax. This is the default when configured for any
10213 PDP-11 target other than @samp{pdp11-*-bsd}.
10216 @node PowerPC Options
10217 @subsection PowerPC Options
10218 @cindex PowerPC options
10220 These are listed under @xref{RS/6000 and PowerPC Options}.
10222 @node RS/6000 and PowerPC Options
10223 @subsection IBM RS/6000 and PowerPC Options
10224 @cindex RS/6000 and PowerPC Options
10225 @cindex IBM RS/6000 and PowerPC Options
10227 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
10234 @itemx -mno-powerpc
10235 @itemx -mpowerpc-gpopt
10236 @itemx -mno-powerpc-gpopt
10237 @itemx -mpowerpc-gfxopt
10238 @itemx -mno-powerpc-gfxopt
10240 @itemx -mno-powerpc64
10244 @opindex mno-power2
10246 @opindex mno-powerpc
10247 @opindex mpowerpc-gpopt
10248 @opindex mno-powerpc-gpopt
10249 @opindex mpowerpc-gfxopt
10250 @opindex mno-powerpc-gfxopt
10251 @opindex mpowerpc64
10252 @opindex mno-powerpc64
10253 GCC supports two related instruction set architectures for the
10254 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
10255 instructions supported by the @samp{rios} chip set used in the original
10256 RS/6000 systems and the @dfn{PowerPC} instruction set is the
10257 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
10258 the IBM 4xx microprocessors.
10260 Neither architecture is a subset of the other. However there is a
10261 large common subset of instructions supported by both. An MQ
10262 register is included in processors supporting the POWER architecture.
10264 You use these options to specify which instructions are available on the
10265 processor you are using. The default value of these options is
10266 determined when configuring GCC@. Specifying the
10267 @option{-mcpu=@var{cpu_type}} overrides the specification of these
10268 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
10269 rather than the options listed above.
10271 The @option{-mpower} option allows GCC to generate instructions that
10272 are found only in the POWER architecture and to use the MQ register.
10273 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
10274 to generate instructions that are present in the POWER2 architecture but
10275 not the original POWER architecture.
10277 The @option{-mpowerpc} option allows GCC to generate instructions that
10278 are found only in the 32-bit subset of the PowerPC architecture.
10279 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
10280 GCC to use the optional PowerPC architecture instructions in the
10281 General Purpose group, including floating-point square root. Specifying
10282 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
10283 use the optional PowerPC architecture instructions in the Graphics
10284 group, including floating-point select.
10286 The @option{-mpowerpc64} option allows GCC to generate the additional
10287 64-bit instructions that are found in the full PowerPC64 architecture
10288 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
10289 @option{-mno-powerpc64}.
10291 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
10292 will use only the instructions in the common subset of both
10293 architectures plus some special AIX common-mode calls, and will not use
10294 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
10295 permits GCC to use any instruction from either architecture and to
10296 allow use of the MQ register; specify this for the Motorola MPC601.
10298 @item -mnew-mnemonics
10299 @itemx -mold-mnemonics
10300 @opindex mnew-mnemonics
10301 @opindex mold-mnemonics
10302 Select which mnemonics to use in the generated assembler code. With
10303 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
10304 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
10305 assembler mnemonics defined for the POWER architecture. Instructions
10306 defined in only one architecture have only one mnemonic; GCC uses that
10307 mnemonic irrespective of which of these options is specified.
10309 GCC defaults to the mnemonics appropriate for the architecture in
10310 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
10311 value of these option. Unless you are building a cross-compiler, you
10312 should normally not specify either @option{-mnew-mnemonics} or
10313 @option{-mold-mnemonics}, but should instead accept the default.
10315 @item -mcpu=@var{cpu_type}
10317 Set architecture type, register usage, choice of mnemonics, and
10318 instruction scheduling parameters for machine type @var{cpu_type}.
10319 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
10320 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505},
10321 @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
10322 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
10323 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
10324 @samp{860}, @samp{970}, @samp{common}, @samp{ec603e}, @samp{G3},
10325 @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3},
10326 @samp{power4}, @samp{power5}, @samp{powerpc}, @samp{powerpc64},
10327 @samp{rios}, @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64a}.
10329 @option{-mcpu=common} selects a completely generic processor. Code
10330 generated under this option will run on any POWER or PowerPC processor.
10331 GCC will use only the instructions in the common subset of both
10332 architectures, and will not use the MQ register. GCC assumes a generic
10333 processor model for scheduling purposes.
10335 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
10336 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
10337 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
10338 types, with an appropriate, generic processor model assumed for
10339 scheduling purposes.
10341 The other options specify a specific processor. Code generated under
10342 those options will run best on that processor, and may not run at all on
10345 The @option{-mcpu} options automatically enable or disable the
10346 following options: @option{-maltivec}, @option{-mhard-float},
10347 @option{-mmfcrf}, @option{-mmultiple}, @option{-mnew-mnemonics},
10348 @option{-mpower}, @option{-mpower2}, @option{-mpowerpc64},
10349 @option{-mpowerpc-gpopt}, @option{-mpowerpc-gfxopt},
10350 @option{-mstring}. The particular options set for any particular CPU
10351 will vary between compiler versions, depending on what setting seems
10352 to produce optimal code for that CPU; it doesn't necessarily reflect
10353 the actual hardware's capabilities. If you wish to set an individual
10354 option to a particular value, you may specify it after the
10355 @option{-mcpu} option, like @samp{-mcpu=970 -mno-altivec}.
10357 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
10358 not enabled or disabled by the @option{-mcpu} option at present, since
10359 AIX does not have full support for these options. You may still
10360 enable or disable them individually if you're sure it'll work in your
10363 @item -mtune=@var{cpu_type}
10365 Set the instruction scheduling parameters for machine type
10366 @var{cpu_type}, but do not set the architecture type, register usage, or
10367 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
10368 values for @var{cpu_type} are used for @option{-mtune} as for
10369 @option{-mcpu}. If both are specified, the code generated will use the
10370 architecture, registers, and mnemonics set by @option{-mcpu}, but the
10371 scheduling parameters set by @option{-mtune}.
10374 @itemx -mno-altivec
10376 @opindex mno-altivec
10377 Generate code that uses (does not use) AltiVec instructions, and also
10378 enable the use of built-in functions that allow more direct access to
10379 the AltiVec instruction set. You may also need to set
10380 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
10385 Extend the current ABI with SPE ABI extensions. This does not change
10386 the default ABI, instead it adds the SPE ABI extensions to the current
10390 @opindex mabi=no-spe
10391 Disable Booke SPE ABI extensions for the current ABI@.
10393 @item -misel=@var{yes/no}
10396 This switch enables or disables the generation of ISEL instructions.
10398 @item -mspe=@var{yes/no}
10401 This switch enables or disables the generation of SPE simd
10404 @item -mfloat-gprs=@var{yes/single/double/no}
10405 @itemx -mfloat-gprs
10406 @opindex mfloat-gprs
10407 This switch enables or disables the generation of floating point
10408 operations on the general purpose registers for architectures that
10411 The argument @var{yes} or @var{single} enables the use of
10412 single-precision floating point operations.
10414 The argument @var{double} enables the use of single and
10415 double-precision floating point operations.
10417 The argument @var{no} disables floating point operations on the
10418 general purpose registers.
10420 This option is currently only available on the MPC854x.
10426 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
10427 targets (including GNU/Linux). The 32-bit environment sets int, long
10428 and pointer to 32 bits and generates code that runs on any PowerPC
10429 variant. The 64-bit environment sets int to 32 bits and long and
10430 pointer to 64 bits, and generates code for PowerPC64, as for
10431 @option{-mpowerpc64}.
10434 @itemx -mno-fp-in-toc
10435 @itemx -mno-sum-in-toc
10436 @itemx -mminimal-toc
10438 @opindex mno-fp-in-toc
10439 @opindex mno-sum-in-toc
10440 @opindex mminimal-toc
10441 Modify generation of the TOC (Table Of Contents), which is created for
10442 every executable file. The @option{-mfull-toc} option is selected by
10443 default. In that case, GCC will allocate at least one TOC entry for
10444 each unique non-automatic variable reference in your program. GCC
10445 will also place floating-point constants in the TOC@. However, only
10446 16,384 entries are available in the TOC@.
10448 If you receive a linker error message that saying you have overflowed
10449 the available TOC space, you can reduce the amount of TOC space used
10450 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
10451 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
10452 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
10453 generate code to calculate the sum of an address and a constant at
10454 run-time instead of putting that sum into the TOC@. You may specify one
10455 or both of these options. Each causes GCC to produce very slightly
10456 slower and larger code at the expense of conserving TOC space.
10458 If you still run out of space in the TOC even when you specify both of
10459 these options, specify @option{-mminimal-toc} instead. This option causes
10460 GCC to make only one TOC entry for every file. When you specify this
10461 option, GCC will produce code that is slower and larger but which
10462 uses extremely little TOC space. You may wish to use this option
10463 only on files that contain less frequently executed code.
10469 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
10470 @code{long} type, and the infrastructure needed to support them.
10471 Specifying @option{-maix64} implies @option{-mpowerpc64} and
10472 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
10473 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
10476 @itemx -mno-xl-call
10478 @opindex mno-xl-call
10479 On AIX, pass floating-point arguments to prototyped functions beyond the
10480 register save area (RSA) on the stack in addition to argument FPRs. The
10481 AIX calling convention was extended but not initially documented to
10482 handle an obscure K&R C case of calling a function that takes the
10483 address of its arguments with fewer arguments than declared. AIX XL
10484 compilers access floating point arguments which do not fit in the
10485 RSA from the stack when a subroutine is compiled without
10486 optimization. Because always storing floating-point arguments on the
10487 stack is inefficient and rarely needed, this option is not enabled by
10488 default and only is necessary when calling subroutines compiled by AIX
10489 XL compilers without optimization.
10493 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
10494 application written to use message passing with special startup code to
10495 enable the application to run. The system must have PE installed in the
10496 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
10497 must be overridden with the @option{-specs=} option to specify the
10498 appropriate directory location. The Parallel Environment does not
10499 support threads, so the @option{-mpe} option and the @option{-pthread}
10500 option are incompatible.
10502 @item -malign-natural
10503 @itemx -malign-power
10504 @opindex malign-natural
10505 @opindex malign-power
10506 On AIX, Darwin, and 64-bit PowerPC GNU/Linux, the option
10507 @option{-malign-natural} overrides the ABI-defined alignment of larger
10508 types, such as floating-point doubles, on their natural size-based boundary.
10509 The option @option{-malign-power} instructs GCC to follow the ABI-specified
10510 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
10513 @itemx -mhard-float
10514 @opindex msoft-float
10515 @opindex mhard-float
10516 Generate code that does not use (uses) the floating-point register set.
10517 Software floating point emulation is provided if you use the
10518 @option{-msoft-float} option, and pass the option to GCC when linking.
10521 @itemx -mno-multiple
10523 @opindex mno-multiple
10524 Generate code that uses (does not use) the load multiple word
10525 instructions and the store multiple word instructions. These
10526 instructions are generated by default on POWER systems, and not
10527 generated on PowerPC systems. Do not use @option{-mmultiple} on little
10528 endian PowerPC systems, since those instructions do not work when the
10529 processor is in little endian mode. The exceptions are PPC740 and
10530 PPC750 which permit the instructions usage in little endian mode.
10535 @opindex mno-string
10536 Generate code that uses (does not use) the load string instructions
10537 and the store string word instructions to save multiple registers and
10538 do small block moves. These instructions are generated by default on
10539 POWER systems, and not generated on PowerPC systems. Do not use
10540 @option{-mstring} on little endian PowerPC systems, since those
10541 instructions do not work when the processor is in little endian mode.
10542 The exceptions are PPC740 and PPC750 which permit the instructions
10543 usage in little endian mode.
10548 @opindex mno-update
10549 Generate code that uses (does not use) the load or store instructions
10550 that update the base register to the address of the calculated memory
10551 location. These instructions are generated by default. If you use
10552 @option{-mno-update}, there is a small window between the time that the
10553 stack pointer is updated and the address of the previous frame is
10554 stored, which means code that walks the stack frame across interrupts or
10555 signals may get corrupted data.
10558 @itemx -mno-fused-madd
10559 @opindex mfused-madd
10560 @opindex mno-fused-madd
10561 Generate code that uses (does not use) the floating point multiply and
10562 accumulate instructions. These instructions are generated by default if
10563 hardware floating is used.
10565 @item -mno-bit-align
10567 @opindex mno-bit-align
10568 @opindex mbit-align
10569 On System V.4 and embedded PowerPC systems do not (do) force structures
10570 and unions that contain bit-fields to be aligned to the base type of the
10573 For example, by default a structure containing nothing but 8
10574 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
10575 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
10576 the structure would be aligned to a 1 byte boundary and be one byte in
10579 @item -mno-strict-align
10580 @itemx -mstrict-align
10581 @opindex mno-strict-align
10582 @opindex mstrict-align
10583 On System V.4 and embedded PowerPC systems do not (do) assume that
10584 unaligned memory references will be handled by the system.
10586 @item -mrelocatable
10587 @itemx -mno-relocatable
10588 @opindex mrelocatable
10589 @opindex mno-relocatable
10590 On embedded PowerPC systems generate code that allows (does not allow)
10591 the program to be relocated to a different address at runtime. If you
10592 use @option{-mrelocatable} on any module, all objects linked together must
10593 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
10595 @item -mrelocatable-lib
10596 @itemx -mno-relocatable-lib
10597 @opindex mrelocatable-lib
10598 @opindex mno-relocatable-lib
10599 On embedded PowerPC systems generate code that allows (does not allow)
10600 the program to be relocated to a different address at runtime. Modules
10601 compiled with @option{-mrelocatable-lib} can be linked with either modules
10602 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
10603 with modules compiled with the @option{-mrelocatable} options.
10609 On System V.4 and embedded PowerPC systems do not (do) assume that
10610 register 2 contains a pointer to a global area pointing to the addresses
10611 used in the program.
10614 @itemx -mlittle-endian
10616 @opindex mlittle-endian
10617 On System V.4 and embedded PowerPC systems compile code for the
10618 processor in little endian mode. The @option{-mlittle-endian} option is
10619 the same as @option{-mlittle}.
10622 @itemx -mbig-endian
10624 @opindex mbig-endian
10625 On System V.4 and embedded PowerPC systems compile code for the
10626 processor in big endian mode. The @option{-mbig-endian} option is
10627 the same as @option{-mbig}.
10629 @item -mdynamic-no-pic
10630 @opindex mdynamic-no-pic
10631 On Darwin and Mac OS X systems, compile code so that it is not
10632 relocatable, but that its external references are relocatable. The
10633 resulting code is suitable for applications, but not shared
10636 @item -mprioritize-restricted-insns=@var{priority}
10637 @opindex mprioritize-restricted-insns
10638 This option controls the priority that is assigned to
10639 dispatch-slot restricted instructions during the second scheduling
10640 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
10641 @var{no/highest/second-highest} priority to dispatch slot restricted
10644 @item -msched-costly-dep=@var{dependence_type}
10645 @opindex msched-costly-dep
10646 This option controls which dependences are considered costly
10647 by the target during instruction scheduling. The argument
10648 @var{dependence_type} takes one of the following values:
10649 @var{no}: no dependence is costly,
10650 @var{all}: all dependences are costly,
10651 @var{true_store_to_load}: a true dependence from store to load is costly,
10652 @var{store_to_load}: any dependence from store to load is costly,
10653 @var{number}: any dependence which latency >= @var{number} is costly.
10655 @item -minsert-sched-nops=@var{scheme}
10656 @opindex minsert-sched-nops
10657 This option controls which nop insertion scheme will be used during
10658 the second scheduling pass. The argument @var{scheme} takes one of the
10660 @var{no}: Don't insert nops.
10661 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
10662 according to the scheduler's grouping.
10663 @var{regroup_exact}: Insert nops to force costly dependent insns into
10664 separate groups. Insert exactly as many nops as needed to force an insn
10665 to a new group, according to the estimated processor grouping.
10666 @var{number}: Insert nops to force costly dependent insns into
10667 separate groups. Insert @var{number} nops to force an insn to a new group.
10670 @opindex mcall-sysv
10671 On System V.4 and embedded PowerPC systems compile code using calling
10672 conventions that adheres to the March 1995 draft of the System V
10673 Application Binary Interface, PowerPC processor supplement. This is the
10674 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
10676 @item -mcall-sysv-eabi
10677 @opindex mcall-sysv-eabi
10678 Specify both @option{-mcall-sysv} and @option{-meabi} options.
10680 @item -mcall-sysv-noeabi
10681 @opindex mcall-sysv-noeabi
10682 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
10684 @item -mcall-solaris
10685 @opindex mcall-solaris
10686 On System V.4 and embedded PowerPC systems compile code for the Solaris
10690 @opindex mcall-linux
10691 On System V.4 and embedded PowerPC systems compile code for the
10692 Linux-based GNU system.
10696 On System V.4 and embedded PowerPC systems compile code for the
10697 Hurd-based GNU system.
10699 @item -mcall-netbsd
10700 @opindex mcall-netbsd
10701 On System V.4 and embedded PowerPC systems compile code for the
10702 NetBSD operating system.
10704 @item -maix-struct-return
10705 @opindex maix-struct-return
10706 Return all structures in memory (as specified by the AIX ABI)@.
10708 @item -msvr4-struct-return
10709 @opindex msvr4-struct-return
10710 Return structures smaller than 8 bytes in registers (as specified by the
10713 @item -mabi=altivec
10714 @opindex mabi=altivec
10715 Extend the current ABI with AltiVec ABI extensions. This does not
10716 change the default ABI, instead it adds the AltiVec ABI extensions to
10719 @item -mabi=no-altivec
10720 @opindex mabi=no-altivec
10721 Disable AltiVec ABI extensions for the current ABI@.
10724 @itemx -mno-prototype
10725 @opindex mprototype
10726 @opindex mno-prototype
10727 On System V.4 and embedded PowerPC systems assume that all calls to
10728 variable argument functions are properly prototyped. Otherwise, the
10729 compiler must insert an instruction before every non prototyped call to
10730 set or clear bit 6 of the condition code register (@var{CR}) to
10731 indicate whether floating point values were passed in the floating point
10732 registers in case the function takes a variable arguments. With
10733 @option{-mprototype}, only calls to prototyped variable argument functions
10734 will set or clear the bit.
10738 On embedded PowerPC systems, assume that the startup module is called
10739 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
10740 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
10745 On embedded PowerPC systems, assume that the startup module is called
10746 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
10751 On embedded PowerPC systems, assume that the startup module is called
10752 @file{crt0.o} and the standard C libraries are @file{libads.a} and
10755 @item -myellowknife
10756 @opindex myellowknife
10757 On embedded PowerPC systems, assume that the startup module is called
10758 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
10763 On System V.4 and embedded PowerPC systems, specify that you are
10764 compiling for a VxWorks system.
10768 Specify that you are compiling for the WindISS simulation environment.
10772 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
10773 header to indicate that @samp{eabi} extended relocations are used.
10779 On System V.4 and embedded PowerPC systems do (do not) adhere to the
10780 Embedded Applications Binary Interface (eabi) which is a set of
10781 modifications to the System V.4 specifications. Selecting @option{-meabi}
10782 means that the stack is aligned to an 8 byte boundary, a function
10783 @code{__eabi} is called to from @code{main} to set up the eabi
10784 environment, and the @option{-msdata} option can use both @code{r2} and
10785 @code{r13} to point to two separate small data areas. Selecting
10786 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
10787 do not call an initialization function from @code{main}, and the
10788 @option{-msdata} option will only use @code{r13} to point to a single
10789 small data area. The @option{-meabi} option is on by default if you
10790 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
10793 @opindex msdata=eabi
10794 On System V.4 and embedded PowerPC systems, put small initialized
10795 @code{const} global and static data in the @samp{.sdata2} section, which
10796 is pointed to by register @code{r2}. Put small initialized
10797 non-@code{const} global and static data in the @samp{.sdata} section,
10798 which is pointed to by register @code{r13}. Put small uninitialized
10799 global and static data in the @samp{.sbss} section, which is adjacent to
10800 the @samp{.sdata} section. The @option{-msdata=eabi} option is
10801 incompatible with the @option{-mrelocatable} option. The
10802 @option{-msdata=eabi} option also sets the @option{-memb} option.
10805 @opindex msdata=sysv
10806 On System V.4 and embedded PowerPC systems, put small global and static
10807 data in the @samp{.sdata} section, which is pointed to by register
10808 @code{r13}. Put small uninitialized global and static data in the
10809 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
10810 The @option{-msdata=sysv} option is incompatible with the
10811 @option{-mrelocatable} option.
10813 @item -msdata=default
10815 @opindex msdata=default
10817 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
10818 compile code the same as @option{-msdata=eabi}, otherwise compile code the
10819 same as @option{-msdata=sysv}.
10822 @opindex msdata-data
10823 On System V.4 and embedded PowerPC systems, put small global and static
10824 data in the @samp{.sdata} section. Put small uninitialized global and
10825 static data in the @samp{.sbss} section. Do not use register @code{r13}
10826 to address small data however. This is the default behavior unless
10827 other @option{-msdata} options are used.
10831 @opindex msdata=none
10833 On embedded PowerPC systems, put all initialized global and static data
10834 in the @samp{.data} section, and all uninitialized data in the
10835 @samp{.bss} section.
10839 @cindex smaller data references (PowerPC)
10840 @cindex .sdata/.sdata2 references (PowerPC)
10841 On embedded PowerPC systems, put global and static items less than or
10842 equal to @var{num} bytes into the small data or bss sections instead of
10843 the normal data or bss section. By default, @var{num} is 8. The
10844 @option{-G @var{num}} switch is also passed to the linker.
10845 All modules should be compiled with the same @option{-G @var{num}} value.
10848 @itemx -mno-regnames
10850 @opindex mno-regnames
10851 On System V.4 and embedded PowerPC systems do (do not) emit register
10852 names in the assembly language output using symbolic forms.
10855 @itemx -mno-longcall
10857 @opindex mno-longcall
10858 Default to making all function calls indirectly, using a register, so
10859 that functions which reside further than 32 megabytes (33,554,432
10860 bytes) from the current location can be called. This setting can be
10861 overridden by the @code{shortcall} function attribute, or by
10862 @code{#pragma longcall(0)}.
10864 Some linkers are capable of detecting out-of-range calls and generating
10865 glue code on the fly. On these systems, long calls are unnecessary and
10866 generate slower code. As of this writing, the AIX linker can do this,
10867 as can the GNU linker for PowerPC/64. It is planned to add this feature
10868 to the GNU linker for 32-bit PowerPC systems as well.
10870 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
10871 callee, L42'', plus a ``branch island'' (glue code). The two target
10872 addresses represent the callee and the ``branch island''. The
10873 Darwin/PPC linker will prefer the first address and generate a ``bl
10874 callee'' if the PPC ``bl'' instruction will reach the callee directly;
10875 otherwise, the linker will generate ``bl L42'' to call the ``branch
10876 island''. The ``branch island'' is appended to the body of the
10877 calling function; it computes the full 32-bit address of the callee
10880 On Mach-O (Darwin) systems, this option directs the compiler emit to
10881 the glue for every direct call, and the Darwin linker decides whether
10882 to use or discard it.
10884 In the future, we may cause GCC to ignore all longcall specifications
10885 when the linker is known to generate glue.
10889 Adds support for multithreading with the @dfn{pthreads} library.
10890 This option sets flags for both the preprocessor and linker.
10894 @node S/390 and zSeries Options
10895 @subsection S/390 and zSeries Options
10896 @cindex S/390 and zSeries Options
10898 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
10902 @itemx -msoft-float
10903 @opindex mhard-float
10904 @opindex msoft-float
10905 Use (do not use) the hardware floating-point instructions and registers
10906 for floating-point operations. When @option{-msoft-float} is specified,
10907 functions in @file{libgcc.a} will be used to perform floating-point
10908 operations. When @option{-mhard-float} is specified, the compiler
10909 generates IEEE floating-point instructions. This is the default.
10912 @itemx -mno-backchain
10913 @opindex mbackchain
10914 @opindex mno-backchain
10915 Store (do not store) the address of the caller's frame as backchain pointer
10916 into the callee's stack frame.
10917 A backchain may be needed to allow debugging using tools that do not understand
10918 DWARF-2 call frame information.
10919 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
10920 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
10921 the backchain is placed into the topmost word of the 96/160 byte register
10924 In general, code compiled with @option{-mbackchain} is call-compatible with
10925 code compiled with @option{-mmo-backchain}; however, use of the backchain
10926 for debugging purposes usually requires that the whole binary is built with
10927 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
10928 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
10929 to build a linux kernel use @option{-msoft-float}.
10931 The default is to not maintain the backchain.
10933 @item -mpacked-stack
10934 @item -mno-packed-stack
10935 @opindex mpacked-stack
10936 @opindex mno-packed-stack
10937 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
10938 specified, the compiler uses the all fields of the 96/160 byte register save
10939 area only for their default purpose; unused fields still take up stack space.
10940 When @option{-mpacked-stack} is specified, register save slots are densely
10941 packed at the top of the register save area; unused space is reused for other
10942 purposes, allowing for more efficient use of the available stack space.
10943 However, when @option{-mbackchain} is also in effect, the topmost word of
10944 the save area is always used to store the backchain, and the return address
10945 register is always saved two words below the backchain.
10947 As long as the stack frame backchain is not used, code generated with
10948 @option{-mpacked-stack} is call-compatible with code generated with
10949 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
10950 S/390 or zSeries generated code that uses the stack frame backchain at run
10951 time, not just for debugging purposes. Such code is not call-compatible
10952 with code compiled with @option{-mpacked-stack}. Also, note that the
10953 combination of @option{-mbackchain},
10954 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
10955 to build a linux kernel use @option{-msoft-float}.
10957 The default is to not use the packed stack layout.
10960 @itemx -mno-small-exec
10961 @opindex msmall-exec
10962 @opindex mno-small-exec
10963 Generate (or do not generate) code using the @code{bras} instruction
10964 to do subroutine calls.
10965 This only works reliably if the total executable size does not
10966 exceed 64k. The default is to use the @code{basr} instruction instead,
10967 which does not have this limitation.
10973 When @option{-m31} is specified, generate code compliant to the
10974 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
10975 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
10976 particular to generate 64-bit instructions. For the @samp{s390}
10977 targets, the default is @option{-m31}, while the @samp{s390x}
10978 targets default to @option{-m64}.
10984 When @option{-mzarch} is specified, generate code using the
10985 instructions available on z/Architecture.
10986 When @option{-mesa} is specified, generate code using the
10987 instructions available on ESA/390. Note that @option{-mesa} is
10988 not possible with @option{-m64}.
10989 When generating code compliant to the GNU/Linux for S/390 ABI,
10990 the default is @option{-mesa}. When generating code compliant
10991 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
10997 Generate (or do not generate) code using the @code{mvcle} instruction
10998 to perform block moves. When @option{-mno-mvcle} is specified,
10999 use a @code{mvc} loop instead. This is the default.
11005 Print (or do not print) additional debug information when compiling.
11006 The default is to not print debug information.
11008 @item -march=@var{cpu-type}
11010 Generate code that will run on @var{cpu-type}, which is the name of a system
11011 representing a certain processor type. Possible values for
11012 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
11013 When generating code using the instructions available on z/Architecture,
11014 the default is @option{-march=z900}. Otherwise, the default is
11015 @option{-march=g5}.
11017 @item -mtune=@var{cpu-type}
11019 Tune to @var{cpu-type} everything applicable about the generated code,
11020 except for the ABI and the set of available instructions.
11021 The list of @var{cpu-type} values is the same as for @option{-march}.
11022 The default is the value used for @option{-march}.
11025 @itemx -mno-tpf-trace
11026 @opindex mtpf-trace
11027 @opindex mno-tpf-trace
11028 Generate code that adds (does not add) in TPF OS specific branches to trace
11029 routines in the operating system. This option is off by default, even
11030 when compiling for the TPF OS@.
11033 @itemx -mno-fused-madd
11034 @opindex mfused-madd
11035 @opindex mno-fused-madd
11036 Generate code that uses (does not use) the floating point multiply and
11037 accumulate instructions. These instructions are generated by default if
11038 hardware floating point is used.
11040 @item -mwarn-framesize=@var{framesize}
11041 @opindex mwarn-framesize
11042 Emit a warning if the current function exceeds the given frame size. Because
11043 this is a compile time check it doesn't need to be a real problem when the program
11044 runs. It is intended to identify functions which most probably cause
11045 a stack overflow. It is useful to be used in an environment with limited stack
11046 size e.g.@: the linux kernel.
11048 @item -mwarn-dynamicstack
11049 @opindex mwarn-dynamicstack
11050 Emit a warning if the function calls alloca or uses dynamically
11051 sized arrays. This is generally a bad idea with a limited stack size.
11053 @item -mstack-guard=@var{stack-guard}
11054 @item -mstack-size=@var{stack-size}
11055 @opindex mstack-guard
11056 @opindex mstack-size
11057 These arguments always have to be used in conjunction. If they are present the s390
11058 back end emits additional instructions in the function prologue which trigger a trap
11059 if the stack size is @var{stack-guard} bytes above the @var{stack-size}
11060 (remember that the stack on s390 grows downward). These options are intended to
11061 be used to help debugging stack overflow problems. The additionally emitted code
11062 cause only little overhead and hence can also be used in production like systems
11063 without greater performance degradation. The given values have to be exact
11064 powers of 2 and @var{stack-size} has to be greater than @var{stack-guard}.
11065 In order to be efficient the extra code makes the assumption that the stack starts
11066 at an address aligned to the value given by @var{stack-size}.
11070 @subsection SH Options
11072 These @samp{-m} options are defined for the SH implementations:
11077 Generate code for the SH1.
11081 Generate code for the SH2.
11084 Generate code for the SH2e.
11088 Generate code for the SH3.
11092 Generate code for the SH3e.
11096 Generate code for the SH4 without a floating-point unit.
11098 @item -m4-single-only
11099 @opindex m4-single-only
11100 Generate code for the SH4 with a floating-point unit that only
11101 supports single-precision arithmetic.
11105 Generate code for the SH4 assuming the floating-point unit is in
11106 single-precision mode by default.
11110 Generate code for the SH4.
11114 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
11115 floating-point unit is not used.
11117 @item -m4a-single-only
11118 @opindex m4a-single-only
11119 Generate code for the SH4a, in such a way that no double-precision
11120 floating point operations are used.
11123 @opindex m4a-single
11124 Generate code for the SH4a assuming the floating-point unit is in
11125 single-precision mode by default.
11129 Generate code for the SH4a.
11133 Same as @option{-m4a-nofpu}, except that it implicitly passes
11134 @option{-dsp} to the assembler. GCC doesn't generate any DSP
11135 instructions at the moment.
11139 Compile code for the processor in big endian mode.
11143 Compile code for the processor in little endian mode.
11147 Align doubles at 64-bit boundaries. Note that this changes the calling
11148 conventions, and thus some functions from the standard C library will
11149 not work unless you recompile it first with @option{-mdalign}.
11153 Shorten some address references at link time, when possible; uses the
11154 linker option @option{-relax}.
11158 Use 32-bit offsets in @code{switch} tables. The default is to use
11163 Enable the use of the instruction @code{fmovd}.
11167 Comply with the calling conventions defined by Renesas.
11171 Comply with the calling conventions defined by Renesas.
11175 Comply with the calling conventions defined for GCC before the Renesas
11176 conventions were available. This option is the default for all
11177 targets of the SH toolchain except for @samp{sh-symbianelf}.
11180 @opindex mnomacsave
11181 Mark the @code{MAC} register as call-clobbered, even if
11182 @option{-mhitachi} is given.
11186 Increase IEEE-compliance of floating-point code.
11190 Dump instruction size and location in the assembly code.
11193 @opindex mpadstruct
11194 This option is deprecated. It pads structures to multiple of 4 bytes,
11195 which is incompatible with the SH ABI@.
11199 Optimize for space instead of speed. Implied by @option{-Os}.
11202 @opindex mprefergot
11203 When generating position-independent code, emit function calls using
11204 the Global Offset Table instead of the Procedure Linkage Table.
11208 Generate a library function call to invalidate instruction cache
11209 entries, after fixing up a trampoline. This library function call
11210 doesn't assume it can write to the whole memory address space. This
11211 is the default when the target is @code{sh-*-linux*}.
11214 @node SPARC Options
11215 @subsection SPARC Options
11216 @cindex SPARC options
11218 These @samp{-m} options are supported on the SPARC:
11221 @item -mno-app-regs
11223 @opindex mno-app-regs
11225 Specify @option{-mapp-regs} to generate output using the global registers
11226 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
11229 To be fully SVR4 ABI compliant at the cost of some performance loss,
11230 specify @option{-mno-app-regs}. You should compile libraries and system
11231 software with this option.
11234 @itemx -mhard-float
11236 @opindex mhard-float
11237 Generate output containing floating point instructions. This is the
11241 @itemx -msoft-float
11243 @opindex msoft-float
11244 Generate output containing library calls for floating point.
11245 @strong{Warning:} the requisite libraries are not available for all SPARC
11246 targets. Normally the facilities of the machine's usual C compiler are
11247 used, but this cannot be done directly in cross-compilation. You must make
11248 your own arrangements to provide suitable library functions for
11249 cross-compilation. The embedded targets @samp{sparc-*-aout} and
11250 @samp{sparclite-*-*} do provide software floating point support.
11252 @option{-msoft-float} changes the calling convention in the output file;
11253 therefore, it is only useful if you compile @emph{all} of a program with
11254 this option. In particular, you need to compile @file{libgcc.a}, the
11255 library that comes with GCC, with @option{-msoft-float} in order for
11258 @item -mhard-quad-float
11259 @opindex mhard-quad-float
11260 Generate output containing quad-word (long double) floating point
11263 @item -msoft-quad-float
11264 @opindex msoft-quad-float
11265 Generate output containing library calls for quad-word (long double)
11266 floating point instructions. The functions called are those specified
11267 in the SPARC ABI@. This is the default.
11269 As of this writing, there are no SPARC implementations that have hardware
11270 support for the quad-word floating point instructions. They all invoke
11271 a trap handler for one of these instructions, and then the trap handler
11272 emulates the effect of the instruction. Because of the trap handler overhead,
11273 this is much slower than calling the ABI library routines. Thus the
11274 @option{-msoft-quad-float} option is the default.
11276 @item -mno-unaligned-doubles
11277 @itemx -munaligned-doubles
11278 @opindex mno-unaligned-doubles
11279 @opindex munaligned-doubles
11280 Assume that doubles have 8 byte alignment. This is the default.
11282 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
11283 alignment only if they are contained in another type, or if they have an
11284 absolute address. Otherwise, it assumes they have 4 byte alignment.
11285 Specifying this option avoids some rare compatibility problems with code
11286 generated by other compilers. It is not the default because it results
11287 in a performance loss, especially for floating point code.
11289 @item -mno-faster-structs
11290 @itemx -mfaster-structs
11291 @opindex mno-faster-structs
11292 @opindex mfaster-structs
11293 With @option{-mfaster-structs}, the compiler assumes that structures
11294 should have 8 byte alignment. This enables the use of pairs of
11295 @code{ldd} and @code{std} instructions for copies in structure
11296 assignment, in place of twice as many @code{ld} and @code{st} pairs.
11297 However, the use of this changed alignment directly violates the SPARC
11298 ABI@. Thus, it's intended only for use on targets where the developer
11299 acknowledges that their resulting code will not be directly in line with
11300 the rules of the ABI@.
11302 @item -mimpure-text
11303 @opindex mimpure-text
11304 @option{-mimpure-text}, used in addition to @option{-shared}, tells
11305 the compiler to not pass @option{-z text} to the linker when linking a
11306 shared object. Using this option, you can link position-dependent
11307 code into a shared object.
11309 @option{-mimpure-text} suppresses the ``relocations remain against
11310 allocatable but non-writable sections'' linker error message.
11311 However, the necessary relocations will trigger copy-on-write, and the
11312 shared object is not actually shared across processes. Instead of
11313 using @option{-mimpure-text}, you should compile all source code with
11314 @option{-fpic} or @option{-fPIC}.
11316 This option is only available on SunOS and Solaris.
11318 @item -mcpu=@var{cpu_type}
11320 Set the instruction set, register set, and instruction scheduling parameters
11321 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
11322 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
11323 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
11324 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc}, and
11325 @samp{ultrasparc3}.
11327 Default instruction scheduling parameters are used for values that select
11328 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
11329 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
11331 Here is a list of each supported architecture and their supported
11336 v8: supersparc, hypersparc
11337 sparclite: f930, f934, sparclite86x
11339 v9: ultrasparc, ultrasparc3
11342 By default (unless configured otherwise), GCC generates code for the V7
11343 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
11344 additionally optimizes it for the Cypress CY7C602 chip, as used in the
11345 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
11346 SPARCStation 1, 2, IPX etc.
11348 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
11349 architecture. The only difference from V7 code is that the compiler emits
11350 the integer multiply and integer divide instructions which exist in SPARC-V8
11351 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
11352 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
11355 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
11356 the SPARC architecture. This adds the integer multiply, integer divide step
11357 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
11358 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
11359 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
11360 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
11361 MB86934 chip, which is the more recent SPARClite with FPU@.
11363 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
11364 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
11365 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
11366 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
11367 optimizes it for the TEMIC SPARClet chip.
11369 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
11370 architecture. This adds 64-bit integer and floating-point move instructions,
11371 3 additional floating-point condition code registers and conditional move
11372 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
11373 optimizes it for the Sun UltraSPARC I/II chips. With
11374 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
11375 Sun UltraSPARC III chip.
11377 @item -mtune=@var{cpu_type}
11379 Set the instruction scheduling parameters for machine type
11380 @var{cpu_type}, but do not set the instruction set or register set that the
11381 option @option{-mcpu=@var{cpu_type}} would.
11383 The same values for @option{-mcpu=@var{cpu_type}} can be used for
11384 @option{-mtune=@var{cpu_type}}, but the only useful values are those
11385 that select a particular cpu implementation. Those are @samp{cypress},
11386 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
11387 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc}, and
11388 @samp{ultrasparc3}.
11393 @opindex mno-v8plus
11394 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
11395 difference from the V8 ABI is that the global and out registers are
11396 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
11397 mode for all SPARC-V9 processors.
11403 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
11404 Visual Instruction Set extensions. The default is @option{-mno-vis}.
11407 These @samp{-m} options are supported in addition to the above
11408 on SPARC-V9 processors in 64-bit environments:
11411 @item -mlittle-endian
11412 @opindex mlittle-endian
11413 Generate code for a processor running in little-endian mode. It is only
11414 available for a few configurations and most notably not on Solaris.
11420 Generate code for a 32-bit or 64-bit environment.
11421 The 32-bit environment sets int, long and pointer to 32 bits.
11422 The 64-bit environment sets int to 32 bits and long and pointer
11425 @item -mcmodel=medlow
11426 @opindex mcmodel=medlow
11427 Generate code for the Medium/Low code model: 64-bit addresses, programs
11428 must be linked in the low 32 bits of memory. Programs can be statically
11429 or dynamically linked.
11431 @item -mcmodel=medmid
11432 @opindex mcmodel=medmid
11433 Generate code for the Medium/Middle code model: 64-bit addresses, programs
11434 must be linked in the low 44 bits of memory, the text and data segments must
11435 be less than 2GB in size and the data segment must be located within 2GB of
11438 @item -mcmodel=medany
11439 @opindex mcmodel=medany
11440 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
11441 may be linked anywhere in memory, the text and data segments must be less
11442 than 2GB in size and the data segment must be located within 2GB of the
11445 @item -mcmodel=embmedany
11446 @opindex mcmodel=embmedany
11447 Generate code for the Medium/Anywhere code model for embedded systems:
11448 64-bit addresses, the text and data segments must be less than 2GB in
11449 size, both starting anywhere in memory (determined at link time). The
11450 global register %g4 points to the base of the data segment. Programs
11451 are statically linked and PIC is not supported.
11454 @itemx -mno-stack-bias
11455 @opindex mstack-bias
11456 @opindex mno-stack-bias
11457 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
11458 frame pointer if present, are offset by @minus{}2047 which must be added back
11459 when making stack frame references. This is the default in 64-bit mode.
11460 Otherwise, assume no such offset is present.
11463 These switches are supported in addition to the above on Solaris:
11468 Add support for multithreading using the Solaris threads library. This
11469 option sets flags for both the preprocessor and linker. This option does
11470 not affect the thread safety of object code produced by the compiler or
11471 that of libraries supplied with it.
11475 Add support for multithreading using the POSIX threads library. This
11476 option sets flags for both the preprocessor and linker. This option does
11477 not affect the thread safety of object code produced by the compiler or
11478 that of libraries supplied with it.
11481 @node System V Options
11482 @subsection Options for System V
11484 These additional options are available on System V Release 4 for
11485 compatibility with other compilers on those systems:
11490 Create a shared object.
11491 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
11495 Identify the versions of each tool used by the compiler, in a
11496 @code{.ident} assembler directive in the output.
11500 Refrain from adding @code{.ident} directives to the output file (this is
11503 @item -YP,@var{dirs}
11505 Search the directories @var{dirs}, and no others, for libraries
11506 specified with @option{-l}.
11508 @item -Ym,@var{dir}
11510 Look in the directory @var{dir} to find the M4 preprocessor.
11511 The assembler uses this option.
11512 @c This is supposed to go with a -Yd for predefined M4 macro files, but
11513 @c the generic assembler that comes with Solaris takes just -Ym.
11516 @node TMS320C3x/C4x Options
11517 @subsection TMS320C3x/C4x Options
11518 @cindex TMS320C3x/C4x Options
11520 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
11524 @item -mcpu=@var{cpu_type}
11526 Set the instruction set, register set, and instruction scheduling
11527 parameters for machine type @var{cpu_type}. Supported values for
11528 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
11529 @samp{c44}. The default is @samp{c40} to generate code for the
11534 @itemx -msmall-memory
11536 @opindex mbig-memory
11538 @opindex msmall-memory
11540 Generates code for the big or small memory model. The small memory
11541 model assumed that all data fits into one 64K word page. At run-time
11542 the data page (DP) register must be set to point to the 64K page
11543 containing the .bss and .data program sections. The big memory model is
11544 the default and requires reloading of the DP register for every direct
11551 Allow (disallow) allocation of general integer operands into the block
11552 count register BK@.
11558 Enable (disable) generation of code using decrement and branch,
11559 DBcond(D), instructions. This is enabled by default for the C4x. To be
11560 on the safe side, this is disabled for the C3x, since the maximum
11561 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
11562 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
11563 that it can utilize the decrement and branch instruction, but will give
11564 up if there is more than one memory reference in the loop. Thus a loop
11565 where the loop counter is decremented can generate slightly more
11566 efficient code, in cases where the RPTB instruction cannot be utilized.
11568 @item -mdp-isr-reload
11570 @opindex mdp-isr-reload
11572 Force the DP register to be saved on entry to an interrupt service
11573 routine (ISR), reloaded to point to the data section, and restored on
11574 exit from the ISR@. This should not be required unless someone has
11575 violated the small memory model by modifying the DP register, say within
11582 For the C3x use the 24-bit MPYI instruction for integer multiplies
11583 instead of a library call to guarantee 32-bit results. Note that if one
11584 of the operands is a constant, then the multiplication will be performed
11585 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
11586 then squaring operations are performed inline instead of a library call.
11589 @itemx -mno-fast-fix
11591 @opindex mno-fast-fix
11592 The C3x/C4x FIX instruction to convert a floating point value to an
11593 integer value chooses the nearest integer less than or equal to the
11594 floating point value rather than to the nearest integer. Thus if the
11595 floating point number is negative, the result will be incorrectly
11596 truncated an additional code is necessary to detect and correct this
11597 case. This option can be used to disable generation of the additional
11598 code required to correct the result.
11604 Enable (disable) generation of repeat block sequences using the RPTB
11605 instruction for zero overhead looping. The RPTB construct is only used
11606 for innermost loops that do not call functions or jump across the loop
11607 boundaries. There is no advantage having nested RPTB loops due to the
11608 overhead required to save and restore the RC, RS, and RE registers.
11609 This is enabled by default with @option{-O2}.
11611 @item -mrpts=@var{count}
11615 Enable (disable) the use of the single instruction repeat instruction
11616 RPTS@. If a repeat block contains a single instruction, and the loop
11617 count can be guaranteed to be less than the value @var{count}, GCC will
11618 emit a RPTS instruction instead of a RPTB@. If no value is specified,
11619 then a RPTS will be emitted even if the loop count cannot be determined
11620 at compile time. Note that the repeated instruction following RPTS does
11621 not have to be reloaded from memory each iteration, thus freeing up the
11622 CPU buses for operands. However, since interrupts are blocked by this
11623 instruction, it is disabled by default.
11625 @item -mloop-unsigned
11626 @itemx -mno-loop-unsigned
11627 @opindex mloop-unsigned
11628 @opindex mno-loop-unsigned
11629 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
11630 is @math{2^{31} + 1} since these instructions test if the iteration count is
11631 negative to terminate the loop. If the iteration count is unsigned
11632 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
11633 exceeded. This switch allows an unsigned iteration count.
11637 Try to emit an assembler syntax that the TI assembler (asm30) is happy
11638 with. This also enforces compatibility with the API employed by the TI
11639 C3x C compiler. For example, long doubles are passed as structures
11640 rather than in floating point registers.
11646 Generate code that uses registers (stack) for passing arguments to functions.
11647 By default, arguments are passed in registers where possible rather
11648 than by pushing arguments on to the stack.
11650 @item -mparallel-insns
11651 @itemx -mno-parallel-insns
11652 @opindex mparallel-insns
11653 @opindex mno-parallel-insns
11654 Allow the generation of parallel instructions. This is enabled by
11655 default with @option{-O2}.
11657 @item -mparallel-mpy
11658 @itemx -mno-parallel-mpy
11659 @opindex mparallel-mpy
11660 @opindex mno-parallel-mpy
11661 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
11662 provided @option{-mparallel-insns} is also specified. These instructions have
11663 tight register constraints which can pessimize the code generation
11664 of large functions.
11669 @subsection V850 Options
11670 @cindex V850 Options
11672 These @samp{-m} options are defined for V850 implementations:
11676 @itemx -mno-long-calls
11677 @opindex mlong-calls
11678 @opindex mno-long-calls
11679 Treat all calls as being far away (near). If calls are assumed to be
11680 far away, the compiler will always load the functions address up into a
11681 register, and call indirect through the pointer.
11687 Do not optimize (do optimize) basic blocks that use the same index
11688 pointer 4 or more times to copy pointer into the @code{ep} register, and
11689 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
11690 option is on by default if you optimize.
11692 @item -mno-prolog-function
11693 @itemx -mprolog-function
11694 @opindex mno-prolog-function
11695 @opindex mprolog-function
11696 Do not use (do use) external functions to save and restore registers
11697 at the prologue and epilogue of a function. The external functions
11698 are slower, but use less code space if more than one function saves
11699 the same number of registers. The @option{-mprolog-function} option
11700 is on by default if you optimize.
11704 Try to make the code as small as possible. At present, this just turns
11705 on the @option{-mep} and @option{-mprolog-function} options.
11707 @item -mtda=@var{n}
11709 Put static or global variables whose size is @var{n} bytes or less into
11710 the tiny data area that register @code{ep} points to. The tiny data
11711 area can hold up to 256 bytes in total (128 bytes for byte references).
11713 @item -msda=@var{n}
11715 Put static or global variables whose size is @var{n} bytes or less into
11716 the small data area that register @code{gp} points to. The small data
11717 area can hold up to 64 kilobytes.
11719 @item -mzda=@var{n}
11721 Put static or global variables whose size is @var{n} bytes or less into
11722 the first 32 kilobytes of memory.
11726 Specify that the target processor is the V850.
11729 @opindex mbig-switch
11730 Generate code suitable for big switch tables. Use this option only if
11731 the assembler/linker complain about out of range branches within a switch
11736 This option will cause r2 and r5 to be used in the code generated by
11737 the compiler. This setting is the default.
11739 @item -mno-app-regs
11740 @opindex mno-app-regs
11741 This option will cause r2 and r5 to be treated as fixed registers.
11745 Specify that the target processor is the V850E1. The preprocessor
11746 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
11747 this option is used.
11751 Specify that the target processor is the V850E@. The preprocessor
11752 constant @samp{__v850e__} will be defined if this option is used.
11754 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
11755 are defined then a default target processor will be chosen and the
11756 relevant @samp{__v850*__} preprocessor constant will be defined.
11758 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
11759 defined, regardless of which processor variant is the target.
11761 @item -mdisable-callt
11762 @opindex mdisable-callt
11763 This option will suppress generation of the CALLT instruction for the
11764 v850e and v850e1 flavors of the v850 architecture. The default is
11765 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
11770 @subsection VAX Options
11771 @cindex VAX options
11773 These @samp{-m} options are defined for the VAX:
11778 Do not output certain jump instructions (@code{aobleq} and so on)
11779 that the Unix assembler for the VAX cannot handle across long
11784 Do output those jump instructions, on the assumption that you
11785 will assemble with the GNU assembler.
11789 Output code for g-format floating point numbers instead of d-format.
11792 @node x86-64 Options
11793 @subsection x86-64 Options
11794 @cindex x86-64 options
11796 These are listed under @xref{i386 and x86-64 Options}.
11798 @node Xstormy16 Options
11799 @subsection Xstormy16 Options
11800 @cindex Xstormy16 Options
11802 These options are defined for Xstormy16:
11807 Choose startup files and linker script suitable for the simulator.
11810 @node Xtensa Options
11811 @subsection Xtensa Options
11812 @cindex Xtensa Options
11814 These options are supported for Xtensa targets:
11818 @itemx -mno-const16
11820 @opindex mno-const16
11821 Enable or disable use of @code{CONST16} instructions for loading
11822 constant values. The @code{CONST16} instruction is currently not a
11823 standard option from Tensilica. When enabled, @code{CONST16}
11824 instructions are always used in place of the standard @code{L32R}
11825 instructions. The use of @code{CONST16} is enabled by default only if
11826 the @code{L32R} instruction is not available.
11829 @itemx -mno-fused-madd
11830 @opindex mfused-madd
11831 @opindex mno-fused-madd
11832 Enable or disable use of fused multiply/add and multiply/subtract
11833 instructions in the floating-point option. This has no effect if the
11834 floating-point option is not also enabled. Disabling fused multiply/add
11835 and multiply/subtract instructions forces the compiler to use separate
11836 instructions for the multiply and add/subtract operations. This may be
11837 desirable in some cases where strict IEEE 754-compliant results are
11838 required: the fused multiply add/subtract instructions do not round the
11839 intermediate result, thereby producing results with @emph{more} bits of
11840 precision than specified by the IEEE standard. Disabling fused multiply
11841 add/subtract instructions also ensures that the program output is not
11842 sensitive to the compiler's ability to combine multiply and add/subtract
11845 @item -mtext-section-literals
11846 @itemx -mno-text-section-literals
11847 @opindex mtext-section-literals
11848 @opindex mno-text-section-literals
11849 Control the treatment of literal pools. The default is
11850 @option{-mno-text-section-literals}, which places literals in a separate
11851 section in the output file. This allows the literal pool to be placed
11852 in a data RAM/ROM, and it also allows the linker to combine literal
11853 pools from separate object files to remove redundant literals and
11854 improve code size. With @option{-mtext-section-literals}, the literals
11855 are interspersed in the text section in order to keep them as close as
11856 possible to their references. This may be necessary for large assembly
11859 @item -mtarget-align
11860 @itemx -mno-target-align
11861 @opindex mtarget-align
11862 @opindex mno-target-align
11863 When this option is enabled, GCC instructs the assembler to
11864 automatically align instructions to reduce branch penalties at the
11865 expense of some code density. The assembler attempts to widen density
11866 instructions to align branch targets and the instructions following call
11867 instructions. If there are not enough preceding safe density
11868 instructions to align a target, no widening will be performed. The
11869 default is @option{-mtarget-align}. These options do not affect the
11870 treatment of auto-aligned instructions like @code{LOOP}, which the
11871 assembler will always align, either by widening density instructions or
11872 by inserting no-op instructions.
11875 @itemx -mno-longcalls
11876 @opindex mlongcalls
11877 @opindex mno-longcalls
11878 When this option is enabled, GCC instructs the assembler to translate
11879 direct calls to indirect calls unless it can determine that the target
11880 of a direct call is in the range allowed by the call instruction. This
11881 translation typically occurs for calls to functions in other source
11882 files. Specifically, the assembler translates a direct @code{CALL}
11883 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
11884 The default is @option{-mno-longcalls}. This option should be used in
11885 programs where the call target can potentially be out of range. This
11886 option is implemented in the assembler, not the compiler, so the
11887 assembly code generated by GCC will still show direct call
11888 instructions---look at the disassembled object code to see the actual
11889 instructions. Note that the assembler will use an indirect call for
11890 every cross-file call, not just those that really will be out of range.
11893 @node zSeries Options
11894 @subsection zSeries Options
11895 @cindex zSeries options
11897 These are listed under @xref{S/390 and zSeries Options}.
11899 @node Code Gen Options
11900 @section Options for Code Generation Conventions
11901 @cindex code generation conventions
11902 @cindex options, code generation
11903 @cindex run-time options
11905 These machine-independent options control the interface conventions
11906 used in code generation.
11908 Most of them have both positive and negative forms; the negative form
11909 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
11910 one of the forms is listed---the one which is not the default. You
11911 can figure out the other form by either removing @samp{no-} or adding
11915 @item -fbounds-check
11916 @opindex fbounds-check
11917 For front-ends that support it, generate additional code to check that
11918 indices used to access arrays are within the declared range. This is
11919 currently only supported by the Java and Fortran 77 front-ends, where
11920 this option defaults to true and false respectively.
11924 This option generates traps for signed overflow on addition, subtraction,
11925 multiplication operations.
11929 This option instructs the compiler to assume that signed arithmetic
11930 overflow of addition, subtraction and multiplication wraps around
11931 using twos-complement representation. This flag enables some optimizations
11932 and disables other. This option is enabled by default for the Java
11933 front-end, as required by the Java language specification.
11936 @opindex fexceptions
11937 Enable exception handling. Generates extra code needed to propagate
11938 exceptions. For some targets, this implies GCC will generate frame
11939 unwind information for all functions, which can produce significant data
11940 size overhead, although it does not affect execution. If you do not
11941 specify this option, GCC will enable it by default for languages like
11942 C++ which normally require exception handling, and disable it for
11943 languages like C that do not normally require it. However, you may need
11944 to enable this option when compiling C code that needs to interoperate
11945 properly with exception handlers written in C++. You may also wish to
11946 disable this option if you are compiling older C++ programs that don't
11947 use exception handling.
11949 @item -fnon-call-exceptions
11950 @opindex fnon-call-exceptions
11951 Generate code that allows trapping instructions to throw exceptions.
11952 Note that this requires platform-specific runtime support that does
11953 not exist everywhere. Moreover, it only allows @emph{trapping}
11954 instructions to throw exceptions, i.e.@: memory references or floating
11955 point instructions. It does not allow exceptions to be thrown from
11956 arbitrary signal handlers such as @code{SIGALRM}.
11958 @item -funwind-tables
11959 @opindex funwind-tables
11960 Similar to @option{-fexceptions}, except that it will just generate any needed
11961 static data, but will not affect the generated code in any other way.
11962 You will normally not enable this option; instead, a language processor
11963 that needs this handling would enable it on your behalf.
11965 @item -fasynchronous-unwind-tables
11966 @opindex fasynchronous-unwind-tables
11967 Generate unwind table in dwarf2 format, if supported by target machine. The
11968 table is exact at each instruction boundary, so it can be used for stack
11969 unwinding from asynchronous events (such as debugger or garbage collector).
11971 @item -fpcc-struct-return
11972 @opindex fpcc-struct-return
11973 Return ``short'' @code{struct} and @code{union} values in memory like
11974 longer ones, rather than in registers. This convention is less
11975 efficient, but it has the advantage of allowing intercallability between
11976 GCC-compiled files and files compiled with other compilers, particularly
11977 the Portable C Compiler (pcc).
11979 The precise convention for returning structures in memory depends
11980 on the target configuration macros.
11982 Short structures and unions are those whose size and alignment match
11983 that of some integer type.
11985 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
11986 switch is not binary compatible with code compiled with the
11987 @option{-freg-struct-return} switch.
11988 Use it to conform to a non-default application binary interface.
11990 @item -freg-struct-return
11991 @opindex freg-struct-return
11992 Return @code{struct} and @code{union} values in registers when possible.
11993 This is more efficient for small structures than
11994 @option{-fpcc-struct-return}.
11996 If you specify neither @option{-fpcc-struct-return} nor
11997 @option{-freg-struct-return}, GCC defaults to whichever convention is
11998 standard for the target. If there is no standard convention, GCC
11999 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
12000 the principal compiler. In those cases, we can choose the standard, and
12001 we chose the more efficient register return alternative.
12003 @strong{Warning:} code compiled with the @option{-freg-struct-return}
12004 switch is not binary compatible with code compiled with the
12005 @option{-fpcc-struct-return} switch.
12006 Use it to conform to a non-default application binary interface.
12008 @item -fshort-enums
12009 @opindex fshort-enums
12010 Allocate to an @code{enum} type only as many bytes as it needs for the
12011 declared range of possible values. Specifically, the @code{enum} type
12012 will be equivalent to the smallest integer type which has enough room.
12014 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
12015 code that is not binary compatible with code generated without that switch.
12016 Use it to conform to a non-default application binary interface.
12018 @item -fshort-double
12019 @opindex fshort-double
12020 Use the same size for @code{double} as for @code{float}.
12022 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
12023 code that is not binary compatible with code generated without that switch.
12024 Use it to conform to a non-default application binary interface.
12026 @item -fshort-wchar
12027 @opindex fshort-wchar
12028 Override the underlying type for @samp{wchar_t} to be @samp{short
12029 unsigned int} instead of the default for the target. This option is
12030 useful for building programs to run under WINE@.
12032 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
12033 code that is not binary compatible with code generated without that switch.
12034 Use it to conform to a non-default application binary interface.
12036 @item -fshared-data
12037 @opindex fshared-data
12038 Requests that the data and non-@code{const} variables of this
12039 compilation be shared data rather than private data. The distinction
12040 makes sense only on certain operating systems, where shared data is
12041 shared between processes running the same program, while private data
12042 exists in one copy per process.
12045 @opindex fno-common
12046 In C, allocate even uninitialized global variables in the data section of the
12047 object file, rather than generating them as common blocks. This has the
12048 effect that if the same variable is declared (without @code{extern}) in
12049 two different compilations, you will get an error when you link them.
12050 The only reason this might be useful is if you wish to verify that the
12051 program will work on other systems which always work this way.
12055 Ignore the @samp{#ident} directive.
12057 @item -finhibit-size-directive
12058 @opindex finhibit-size-directive
12059 Don't output a @code{.size} assembler directive, or anything else that
12060 would cause trouble if the function is split in the middle, and the
12061 two halves are placed at locations far apart in memory. This option is
12062 used when compiling @file{crtstuff.c}; you should not need to use it
12065 @item -fverbose-asm
12066 @opindex fverbose-asm
12067 Put extra commentary information in the generated assembly code to
12068 make it more readable. This option is generally only of use to those
12069 who actually need to read the generated assembly code (perhaps while
12070 debugging the compiler itself).
12072 @option{-fno-verbose-asm}, the default, causes the
12073 extra information to be omitted and is useful when comparing two assembler
12078 @cindex global offset table
12080 Generate position-independent code (PIC) suitable for use in a shared
12081 library, if supported for the target machine. Such code accesses all
12082 constant addresses through a global offset table (GOT)@. The dynamic
12083 loader resolves the GOT entries when the program starts (the dynamic
12084 loader is not part of GCC; it is part of the operating system). If
12085 the GOT size for the linked executable exceeds a machine-specific
12086 maximum size, you get an error message from the linker indicating that
12087 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
12088 instead. (These maximums are 8k on the SPARC and 32k
12089 on the m68k and RS/6000. The 386 has no such limit.)
12091 Position-independent code requires special support, and therefore works
12092 only on certain machines. For the 386, GCC supports PIC for System V
12093 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
12094 position-independent.
12098 If supported for the target machine, emit position-independent code,
12099 suitable for dynamic linking and avoiding any limit on the size of the
12100 global offset table. This option makes a difference on the m68k,
12101 PowerPC and SPARC@.
12103 Position-independent code requires special support, and therefore works
12104 only on certain machines.
12110 These options are similar to @option{-fpic} and @option{-fPIC}, but
12111 generated position independent code can be only linked into executables.
12112 Usually these options are used when @option{-pie} GCC option will be
12113 used during linking.
12115 @item -ffixed-@var{reg}
12117 Treat the register named @var{reg} as a fixed register; generated code
12118 should never refer to it (except perhaps as a stack pointer, frame
12119 pointer or in some other fixed role).
12121 @var{reg} must be the name of a register. The register names accepted
12122 are machine-specific and are defined in the @code{REGISTER_NAMES}
12123 macro in the machine description macro file.
12125 This flag does not have a negative form, because it specifies a
12128 @item -fcall-used-@var{reg}
12129 @opindex fcall-used
12130 Treat the register named @var{reg} as an allocable register that is
12131 clobbered by function calls. It may be allocated for temporaries or
12132 variables that do not live across a call. Functions compiled this way
12133 will not save and restore the register @var{reg}.
12135 It is an error to used this flag with the frame pointer or stack pointer.
12136 Use of this flag for other registers that have fixed pervasive roles in
12137 the machine's execution model will produce disastrous results.
12139 This flag does not have a negative form, because it specifies a
12142 @item -fcall-saved-@var{reg}
12143 @opindex fcall-saved
12144 Treat the register named @var{reg} as an allocable register saved by
12145 functions. It may be allocated even for temporaries or variables that
12146 live across a call. Functions compiled this way will save and restore
12147 the register @var{reg} if they use it.
12149 It is an error to used this flag with the frame pointer or stack pointer.
12150 Use of this flag for other registers that have fixed pervasive roles in
12151 the machine's execution model will produce disastrous results.
12153 A different sort of disaster will result from the use of this flag for
12154 a register in which function values may be returned.
12156 This flag does not have a negative form, because it specifies a
12159 @item -fpack-struct[=@var{n}]
12160 @opindex fpack-struct
12161 Without a value specified, pack all structure members together without
12162 holes. When a value is specified (which must be a small power of two), pack
12163 structure members according to this value, representing the maximum
12164 alignment (that is, objects with default alignment requirements larger than
12165 this will be output potentially unaligned at the next fitting location.
12167 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
12168 code that is not binary compatible with code generated without that switch.
12169 Additionally, it makes the code suboptimal.
12170 Use it to conform to a non-default application binary interface.
12172 @item -finstrument-functions
12173 @opindex finstrument-functions
12174 Generate instrumentation calls for entry and exit to functions. Just
12175 after function entry and just before function exit, the following
12176 profiling functions will be called with the address of the current
12177 function and its call site. (On some platforms,
12178 @code{__builtin_return_address} does not work beyond the current
12179 function, so the call site information may not be available to the
12180 profiling functions otherwise.)
12183 void __cyg_profile_func_enter (void *this_fn,
12185 void __cyg_profile_func_exit (void *this_fn,
12189 The first argument is the address of the start of the current function,
12190 which may be looked up exactly in the symbol table.
12192 This instrumentation is also done for functions expanded inline in other
12193 functions. The profiling calls will indicate where, conceptually, the
12194 inline function is entered and exited. This means that addressable
12195 versions of such functions must be available. If all your uses of a
12196 function are expanded inline, this may mean an additional expansion of
12197 code size. If you use @samp{extern inline} in your C code, an
12198 addressable version of such functions must be provided. (This is
12199 normally the case anyways, but if you get lucky and the optimizer always
12200 expands the functions inline, you might have gotten away without
12201 providing static copies.)
12203 A function may be given the attribute @code{no_instrument_function}, in
12204 which case this instrumentation will not be done. This can be used, for
12205 example, for the profiling functions listed above, high-priority
12206 interrupt routines, and any functions from which the profiling functions
12207 cannot safely be called (perhaps signal handlers, if the profiling
12208 routines generate output or allocate memory).
12210 @item -fstack-check
12211 @opindex fstack-check
12212 Generate code to verify that you do not go beyond the boundary of the
12213 stack. You should specify this flag if you are running in an
12214 environment with multiple threads, but only rarely need to specify it in
12215 a single-threaded environment since stack overflow is automatically
12216 detected on nearly all systems if there is only one stack.
12218 Note that this switch does not actually cause checking to be done; the
12219 operating system must do that. The switch causes generation of code
12220 to ensure that the operating system sees the stack being extended.
12222 @item -fstack-limit-register=@var{reg}
12223 @itemx -fstack-limit-symbol=@var{sym}
12224 @itemx -fno-stack-limit
12225 @opindex fstack-limit-register
12226 @opindex fstack-limit-symbol
12227 @opindex fno-stack-limit
12228 Generate code to ensure that the stack does not grow beyond a certain value,
12229 either the value of a register or the address of a symbol. If the stack
12230 would grow beyond the value, a signal is raised. For most targets,
12231 the signal is raised before the stack overruns the boundary, so
12232 it is possible to catch the signal without taking special precautions.
12234 For instance, if the stack starts at absolute address @samp{0x80000000}
12235 and grows downwards, you can use the flags
12236 @option{-fstack-limit-symbol=__stack_limit} and
12237 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
12238 of 128KB@. Note that this may only work with the GNU linker.
12240 @cindex aliasing of parameters
12241 @cindex parameters, aliased
12242 @item -fargument-alias
12243 @itemx -fargument-noalias
12244 @itemx -fargument-noalias-global
12245 @opindex fargument-alias
12246 @opindex fargument-noalias
12247 @opindex fargument-noalias-global
12248 Specify the possible relationships among parameters and between
12249 parameters and global data.
12251 @option{-fargument-alias} specifies that arguments (parameters) may
12252 alias each other and may alias global storage.@*
12253 @option{-fargument-noalias} specifies that arguments do not alias
12254 each other, but may alias global storage.@*
12255 @option{-fargument-noalias-global} specifies that arguments do not
12256 alias each other and do not alias global storage.
12258 Each language will automatically use whatever option is required by
12259 the language standard. You should not need to use these options yourself.
12261 @item -fleading-underscore
12262 @opindex fleading-underscore
12263 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
12264 change the way C symbols are represented in the object file. One use
12265 is to help link with legacy assembly code.
12267 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
12268 generate code that is not binary compatible with code generated without that
12269 switch. Use it to conform to a non-default application binary interface.
12270 Not all targets provide complete support for this switch.
12272 @item -ftls-model=@var{model}
12273 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
12274 The @var{model} argument should be one of @code{global-dynamic},
12275 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
12277 The default without @option{-fpic} is @code{initial-exec}; with
12278 @option{-fpic} the default is @code{global-dynamic}.
12280 @item -fvisibility=@var{default|internal|hidden|protected}
12281 @opindex fvisibility
12282 Set the default ELF image symbol visibility to the specified option---all
12283 symbols will be marked with this unless overridden within the code.
12284 Using this feature can very substantially improve linking and
12285 load times of shared object libraries, produce more optimized
12286 code, provide near-perfect API export and prevent symbol clashes.
12287 It is @strong{strongly} recommended that you use this in any shared objects
12290 Despite the nomenclature, @code{default} always means public ie;
12291 available to be linked against from outside the shared object.
12292 @code{protected} and @code{internal} are pretty useless in real-world
12293 usage so the only other commonly used option will be @code{hidden}.
12294 The default if @option{-fvisibility} isn't specified is
12295 @code{default}, i.e., make every
12296 symbol public---this causes the same behavior as previous versions of
12299 A good explanation of the benefits offered by ensuring ELF
12300 symbols have the correct visibility is given by ``How To Write
12301 Shared Libraries'' by Ulrich Drepper (which can be found at
12302 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
12303 solution made possible by this option to marking things hidden when
12304 the default is public is to make the default hidden and mark things
12305 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
12306 and @code{__attribute__ ((visibility("default")))} instead of
12307 @code{__declspec(dllexport)} you get almost identical semantics with
12308 identical syntax. This is a great boon to those working with
12309 cross-platform projects.
12311 For those adding visibility support to existing code, you may find
12312 @samp{#pragma GCC visibility} of use. This works by you enclosing
12313 the declarations you wish to set visibility for with (for example)
12314 @samp{#pragma GCC visibility push(hidden)} and
12315 @samp{#pragma GCC visibility pop}. These can be nested up to sixteen
12316 times. Bear in mind that symbol visibility should be viewed @strong{as
12317 part of the API interface contract} and thus all new code should
12318 always specify visibility when it is not the default ie; declarations
12319 only for use within the local DSO should @strong{always} be marked explicitly
12320 as hidden as so to avoid PLT indirection overheads---making this
12321 abundantly clear also aids readability and self-documentation of the code.
12322 Note that due to ISO C++ specification requirements, operator new and
12323 operator delete must always be of default visibility.
12325 An overview of these techniques, their benefits and how to use them
12326 is at @w{@uref{http://www.nedprod.com/programs/gccvisibility.html}}.
12332 @node Environment Variables
12333 @section Environment Variables Affecting GCC
12334 @cindex environment variables
12336 @c man begin ENVIRONMENT
12337 This section describes several environment variables that affect how GCC
12338 operates. Some of them work by specifying directories or prefixes to use
12339 when searching for various kinds of files. Some are used to specify other
12340 aspects of the compilation environment.
12342 Note that you can also specify places to search using options such as
12343 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
12344 take precedence over places specified using environment variables, which
12345 in turn take precedence over those specified by the configuration of GCC@.
12346 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
12347 GNU Compiler Collection (GCC) Internals}.
12352 @c @itemx LC_COLLATE
12354 @c @itemx LC_MONETARY
12355 @c @itemx LC_NUMERIC
12360 @c @findex LC_COLLATE
12361 @findex LC_MESSAGES
12362 @c @findex LC_MONETARY
12363 @c @findex LC_NUMERIC
12367 These environment variables control the way that GCC uses
12368 localization information that allow GCC to work with different
12369 national conventions. GCC inspects the locale categories
12370 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
12371 so. These locale categories can be set to any value supported by your
12372 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
12373 Kingdom encoded in UTF-8.
12375 The @env{LC_CTYPE} environment variable specifies character
12376 classification. GCC uses it to determine the character boundaries in
12377 a string; this is needed for some multibyte encodings that contain quote
12378 and escape characters that would otherwise be interpreted as a string
12381 The @env{LC_MESSAGES} environment variable specifies the language to
12382 use in diagnostic messages.
12384 If the @env{LC_ALL} environment variable is set, it overrides the value
12385 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
12386 and @env{LC_MESSAGES} default to the value of the @env{LANG}
12387 environment variable. If none of these variables are set, GCC
12388 defaults to traditional C English behavior.
12392 If @env{TMPDIR} is set, it specifies the directory to use for temporary
12393 files. GCC uses temporary files to hold the output of one stage of
12394 compilation which is to be used as input to the next stage: for example,
12395 the output of the preprocessor, which is the input to the compiler
12398 @item GCC_EXEC_PREFIX
12399 @findex GCC_EXEC_PREFIX
12400 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
12401 names of the subprograms executed by the compiler. No slash is added
12402 when this prefix is combined with the name of a subprogram, but you can
12403 specify a prefix that ends with a slash if you wish.
12405 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
12406 an appropriate prefix to use based on the pathname it was invoked with.
12408 If GCC cannot find the subprogram using the specified prefix, it
12409 tries looking in the usual places for the subprogram.
12411 The default value of @env{GCC_EXEC_PREFIX} is
12412 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the value
12413 of @code{prefix} when you ran the @file{configure} script.
12415 Other prefixes specified with @option{-B} take precedence over this prefix.
12417 This prefix is also used for finding files such as @file{crt0.o} that are
12420 In addition, the prefix is used in an unusual way in finding the
12421 directories to search for header files. For each of the standard
12422 directories whose name normally begins with @samp{/usr/local/lib/gcc}
12423 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
12424 replacing that beginning with the specified prefix to produce an
12425 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
12426 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
12427 These alternate directories are searched first; the standard directories
12430 @item COMPILER_PATH
12431 @findex COMPILER_PATH
12432 The value of @env{COMPILER_PATH} is a colon-separated list of
12433 directories, much like @env{PATH}. GCC tries the directories thus
12434 specified when searching for subprograms, if it can't find the
12435 subprograms using @env{GCC_EXEC_PREFIX}.
12438 @findex LIBRARY_PATH
12439 The value of @env{LIBRARY_PATH} is a colon-separated list of
12440 directories, much like @env{PATH}. When configured as a native compiler,
12441 GCC tries the directories thus specified when searching for special
12442 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
12443 using GCC also uses these directories when searching for ordinary
12444 libraries for the @option{-l} option (but directories specified with
12445 @option{-L} come first).
12449 @cindex locale definition
12450 This variable is used to pass locale information to the compiler. One way in
12451 which this information is used is to determine the character set to be used
12452 when character literals, string literals and comments are parsed in C and C++.
12453 When the compiler is configured to allow multibyte characters,
12454 the following values for @env{LANG} are recognized:
12458 Recognize JIS characters.
12460 Recognize SJIS characters.
12462 Recognize EUCJP characters.
12465 If @env{LANG} is not defined, or if it has some other value, then the
12466 compiler will use mblen and mbtowc as defined by the default locale to
12467 recognize and translate multibyte characters.
12471 Some additional environments variables affect the behavior of the
12474 @include cppenv.texi
12478 @node Precompiled Headers
12479 @section Using Precompiled Headers
12480 @cindex precompiled headers
12481 @cindex speed of compilation
12483 Often large projects have many header files that are included in every
12484 source file. The time the compiler takes to process these header files
12485 over and over again can account for nearly all of the time required to
12486 build the project. To make builds faster, GCC allows users to
12487 `precompile' a header file; then, if builds can use the precompiled
12488 header file they will be much faster.
12490 @strong{Caution:} There are a few known situations where GCC will
12491 crash when trying to use a precompiled header. If you have trouble
12492 with a precompiled header, you should remove the precompiled header
12493 and compile without it. In addition, please use GCC's on-line
12494 defect-tracking system to report any problems you encounter with
12495 precompiled headers. @xref{Bugs}.
12497 To create a precompiled header file, simply compile it as you would any
12498 other file, if necessary using the @option{-x} option to make the driver
12499 treat it as a C or C++ header file. You will probably want to use a
12500 tool like @command{make} to keep the precompiled header up-to-date when
12501 the headers it contains change.
12503 A precompiled header file will be searched for when @code{#include} is
12504 seen in the compilation. As it searches for the included file
12505 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
12506 compiler looks for a precompiled header in each directory just before it
12507 looks for the include file in that directory. The name searched for is
12508 the name specified in the @code{#include} with @samp{.gch} appended. If
12509 the precompiled header file can't be used, it is ignored.
12511 For instance, if you have @code{#include "all.h"}, and you have
12512 @file{all.h.gch} in the same directory as @file{all.h}, then the
12513 precompiled header file will be used if possible, and the original
12514 header will be used otherwise.
12516 Alternatively, you might decide to put the precompiled header file in a
12517 directory and use @option{-I} to ensure that directory is searched
12518 before (or instead of) the directory containing the original header.
12519 Then, if you want to check that the precompiled header file is always
12520 used, you can put a file of the same name as the original header in this
12521 directory containing an @code{#error} command.
12523 This also works with @option{-include}. So yet another way to use
12524 precompiled headers, good for projects not designed with precompiled
12525 header files in mind, is to simply take most of the header files used by
12526 a project, include them from another header file, precompile that header
12527 file, and @option{-include} the precompiled header. If the header files
12528 have guards against multiple inclusion, they will be skipped because
12529 they've already been included (in the precompiled header).
12531 If you need to precompile the same header file for different
12532 languages, targets, or compiler options, you can instead make a
12533 @emph{directory} named like @file{all.h.gch}, and put each precompiled
12534 header in the directory, perhaps using @option{-o}. It doesn't matter
12535 what you call the files in the directory, every precompiled header in
12536 the directory will be considered. The first precompiled header
12537 encountered in the directory that is valid for this compilation will
12538 be used; they're searched in no particular order.
12540 There are many other possibilities, limited only by your imagination,
12541 good sense, and the constraints of your build system.
12543 A precompiled header file can be used only when these conditions apply:
12547 Only one precompiled header can be used in a particular compilation.
12550 A precompiled header can't be used once the first C token is seen. You
12551 can have preprocessor directives before a precompiled header; you can
12552 even include a precompiled header from inside another header, so long as
12553 there are no C tokens before the @code{#include}.
12556 The precompiled header file must be produced for the same language as
12557 the current compilation. You can't use a C precompiled header for a C++
12561 The precompiled header file must be produced by the same compiler
12562 version and configuration as the current compilation is using.
12563 The easiest way to guarantee this is to use the same compiler binary
12564 for creating and using precompiled headers.
12567 Any macros defined before the precompiled header is included must
12568 either be defined in the same way as when the precompiled header was
12569 generated, or must not affect the precompiled header, which usually
12570 means that the they don't appear in the precompiled header at all.
12572 The @option{-D} option is one way to define a macro before a
12573 precompiled header is included; using a @code{#define} can also do it.
12574 There are also some options that define macros implicitly, like
12575 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
12578 @item If debugging information is output when using the precompiled
12579 header, using @option{-g} or similar, the same kind of debugging information
12580 must have been output when building the precompiled header. However,
12581 a precompiled header built using @option{-g} can be used in a compilation
12582 when no debugging information is being output.
12584 @item The same @option{-m} options must generally be used when building
12585 and using the precompiled header. @xref{Submodel Options},
12586 for any cases where this rule is relaxed.
12588 @item Each of the following options must be the same when building and using
12589 the precompiled header:
12591 @gccoptlist{-fexceptions -funit-at-a-time}
12594 Some other command-line options starting with @option{-f},
12595 @option{-p}, or @option{-O} must be defined in the same way as when
12596 the precompiled header was generated. At present, it's not clear
12597 which options are safe to change and which are not; the safest choice
12598 is to use exactly the same options when generating and using the
12599 precompiled header. The following are known to be safe:
12601 @gccoptlist{-fpreprocessed -pedantic-errors}
12605 For all of these except the last, the compiler will automatically
12606 ignore the precompiled header if the conditions aren't met. If you
12607 find an option combination that doesn't work and doesn't cause the
12608 precompiled header to be ignored, please consider filing a bug report,
12611 If you do use differing options when generating and using the
12612 precompiled header, the actual behavior will be a mixture of the
12613 behavior for the options. For instance, if you use @option{-g} to
12614 generate the precompiled header but not when using it, you may or may
12615 not get debugging information for routines in the precompiled header.
12617 @node Running Protoize
12618 @section Running Protoize
12620 The program @code{protoize} is an optional part of GCC@. You can use
12621 it to add prototypes to a program, thus converting the program to ISO
12622 C in one respect. The companion program @code{unprotoize} does the
12623 reverse: it removes argument types from any prototypes that are found.
12625 When you run these programs, you must specify a set of source files as
12626 command line arguments. The conversion programs start out by compiling
12627 these files to see what functions they define. The information gathered
12628 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
12630 After scanning comes actual conversion. The specified files are all
12631 eligible to be converted; any files they include (whether sources or
12632 just headers) are eligible as well.
12634 But not all the eligible files are converted. By default,
12635 @code{protoize} and @code{unprotoize} convert only source and header
12636 files in the current directory. You can specify additional directories
12637 whose files should be converted with the @option{-d @var{directory}}
12638 option. You can also specify particular files to exclude with the
12639 @option{-x @var{file}} option. A file is converted if it is eligible, its
12640 directory name matches one of the specified directory names, and its
12641 name within the directory has not been excluded.
12643 Basic conversion with @code{protoize} consists of rewriting most
12644 function definitions and function declarations to specify the types of
12645 the arguments. The only ones not rewritten are those for varargs
12648 @code{protoize} optionally inserts prototype declarations at the
12649 beginning of the source file, to make them available for any calls that
12650 precede the function's definition. Or it can insert prototype
12651 declarations with block scope in the blocks where undeclared functions
12654 Basic conversion with @code{unprotoize} consists of rewriting most
12655 function declarations to remove any argument types, and rewriting
12656 function definitions to the old-style pre-ISO form.
12658 Both conversion programs print a warning for any function declaration or
12659 definition that they can't convert. You can suppress these warnings
12662 The output from @code{protoize} or @code{unprotoize} replaces the
12663 original source file. The original file is renamed to a name ending
12664 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
12665 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
12666 for DOS) file already exists, then the source file is simply discarded.
12668 @code{protoize} and @code{unprotoize} both depend on GCC itself to
12669 scan the program and collect information about the functions it uses.
12670 So neither of these programs will work until GCC is installed.
12672 Here is a table of the options you can use with @code{protoize} and
12673 @code{unprotoize}. Each option works with both programs unless
12677 @item -B @var{directory}
12678 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
12679 usual directory (normally @file{/usr/local/lib}). This file contains
12680 prototype information about standard system functions. This option
12681 applies only to @code{protoize}.
12683 @item -c @var{compilation-options}
12684 Use @var{compilation-options} as the options when running @command{gcc} to
12685 produce the @samp{.X} files. The special option @option{-aux-info} is
12686 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
12688 Note that the compilation options must be given as a single argument to
12689 @code{protoize} or @code{unprotoize}. If you want to specify several
12690 @command{gcc} options, you must quote the entire set of compilation options
12691 to make them a single word in the shell.
12693 There are certain @command{gcc} arguments that you cannot use, because they
12694 would produce the wrong kind of output. These include @option{-g},
12695 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
12696 the @var{compilation-options}, they are ignored.
12699 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
12700 systems) instead of @samp{.c}. This is convenient if you are converting
12701 a C program to C++. This option applies only to @code{protoize}.
12704 Add explicit global declarations. This means inserting explicit
12705 declarations at the beginning of each source file for each function
12706 that is called in the file and was not declared. These declarations
12707 precede the first function definition that contains a call to an
12708 undeclared function. This option applies only to @code{protoize}.
12710 @item -i @var{string}
12711 Indent old-style parameter declarations with the string @var{string}.
12712 This option applies only to @code{protoize}.
12714 @code{unprotoize} converts prototyped function definitions to old-style
12715 function definitions, where the arguments are declared between the
12716 argument list and the initial @samp{@{}. By default, @code{unprotoize}
12717 uses five spaces as the indentation. If you want to indent with just
12718 one space instead, use @option{-i " "}.
12721 Keep the @samp{.X} files. Normally, they are deleted after conversion
12725 Add explicit local declarations. @code{protoize} with @option{-l} inserts
12726 a prototype declaration for each function in each block which calls the
12727 function without any declaration. This option applies only to
12731 Make no real changes. This mode just prints information about the conversions
12732 that would have been done without @option{-n}.
12735 Make no @samp{.save} files. The original files are simply deleted.
12736 Use this option with caution.
12738 @item -p @var{program}
12739 Use the program @var{program} as the compiler. Normally, the name
12740 @file{gcc} is used.
12743 Work quietly. Most warnings are suppressed.
12746 Print the version number, just like @option{-v} for @command{gcc}.
12749 If you need special compiler options to compile one of your program's
12750 source files, then you should generate that file's @samp{.X} file
12751 specially, by running @command{gcc} on that source file with the
12752 appropriate options and the option @option{-aux-info}. Then run
12753 @code{protoize} on the entire set of files. @code{protoize} will use
12754 the existing @samp{.X} file because it is newer than the source file.
12758 gcc -Dfoo=bar file1.c -aux-info file1.X
12763 You need to include the special files along with the rest in the
12764 @code{protoize} command, even though their @samp{.X} files already
12765 exist, because otherwise they won't get converted.
12767 @xref{Protoize Caveats}, for more information on how to use
12768 @code{protoize} successfully.